It probably made things worse, but the fact that they reached out to her to use her voice and she explicitly refused would be sufficient ammo I feel. (Not a lawyer of course).
Of course, Twitter continues to bring people with big egos to their own downfall.
Not to mention that it matches up pretty much exactly with the Bette Midler and Tom Waits cases, where courts ruled against companies using soundalikes after the person they really wanted turned them down. Doesn't matter if they hired a soundalike actress rather than clone her voice, it still violates her rights.
I am guessing this is only because they first tried to hire the originals before hiring sound-alikes... otherwise, would it mean that if your voice sounds similar to someone else, you can't do voice work?
Yes, parody's fine when it's clearly parody. But if you try to pretend that Trump or Obama (rather than an impersonator) is endorsing a product, you're in trouble.
>Wheel of Fortune hostess Vanna White had established herself as a TV personality, and consequently appeared as a spokesperson for advertisers. Samsung produced a television commercial advertising its VCRs, showing a robot wearing a dress and with other similarities to White standing beside a Wheel of Fortune game board. Samsung, in their own internal documents, called this the "Vanna White ad". White sued Samsung for violations of California Civil Code section 3344, California common law right of publicity, and the federal Lanham Act. The United States District Court for the Southern District of California granted summary judgment against White on all counts, and White appealed.
>The Ninth Circuit reversed the District Court, finding that White had a cause of action based on the value of her image, and that Samsung had appropriated this image. Samsung's assertion that this was a parody was found to be unavailing, as the intent of the ad was not to make fun of White's characteristics, but to sell VCRs.
Maybe it depends on which court will handle the case, but OpenAI's core intent isn't parody, but rather to use someone's likeness as a way to make money.
Good voice actors can do a whole range of voices, including imitating many different people. The cases where someone got in trouble are where there's misrepresentation. If it goes to court, there's discovery, and if the OpenAI people gave specific instructions to the voice actor to imitate Scarlett Johansson, after denying it, there could be big trouble. We don't know that, but it looks likely given how they first approached her and how they seemed to be going for something like the "Her" film.
It's only really about the intent of the voice to sound like the original. Reaching out to the originals first implies intent, so it makes the case easier.
It would be harder to find a case if they simply just hired someone that sounds similar, but if they did that with the intention to sound like the original that's still impersonation, only it's harder to prove.
If they just happened to hire someone that sounded like the original, then that's fair game IMO.
Definitely. GPT4o has a voice that sounds like Scarlett Johannson? They'd probably get away with it, I'm sure there are a lot of people that sound like her. Tweeting a reference to a movie she was in - a bit more murky because it's starting to sound like they are deliberately cloning her voice. Asking to use her voice, then using a soundalike, then referencing the movie? 100%, no doubt.
They’ll make the case that the abilities of the device is what he was referring to, but I think more the fact they were pushing her so hard for her involvement will actually be the damning aspect for them with that line of defense.
What are the chances that among 7 billion people in the world that there are always going to be 100 people that sound like you? If Sam Altman was going for a particular voice, there are probably 100 people that indistinguishably have that voice and it just becomes a question of a headhunt.
That's precisely what they did with Doodle God to imitate Morgan Freeman [0] and how James Veich deep faked David Attenborough in his PLnaT eRth video [1].
Trying to imitate her voice to get around paying her wouldn't be okay either. Waits vs. Frito Lay taught us that, if nothing else did. The question is whether people would think they were hearing Scarlett Johansson's voice when using that product, and the answer is yes, so they have to pay her to trade on her identity.
It's not just successful companies though. There is a bit of ego necessary in a founder that makes them think their idea or their implementation of a thing is better so that it needs to be its own company. Sometimes though they even get caught up in their own reality distortion fields with obviously bad ideas or ideas implemented badly due to their own arrogance that ultimately fails.
This is pretty much quintessential founder behavior. I have had my run-ins with people like this, and the relationship is usually short lived. I do not drink the kool-aid, and question pretty much everything. These types of personalities are like oil and water and do not mix. You almost need a third person to act as a emulsifier to allow the oil&water to mix without separating.
Yeah this explains why most of my interviews with startups haven't gone well. I've even had friends ask me to do PoC work for them for equity, and they still get all bent out of shape when I'm not instantly smitten by their one-sentence pitch, and instead ask for more details about their business plan.
If I thought I had a great idea, I would want people to try to poke holes in it. Yet founders often universally seem to be the incredibly sensitive and insecure about their idea.
> "Sam and Jack, I know you remember my Torah portion was about Moses forgiving his brothers. “Forgive them father for they know not what they’ve done” Sexual, physical, emotional, verbal, financial, and technological abuse. Never forgotten."
That is...not a pretty picture. We desperately need someone else at the helm of OpenAI.
Well, this confirms that OpenAI have been shooting from the hip, not that we needed much confirmation. The fact that they repeatedly tried to hire Johansson, then went ahead and made a soundalike while explicitly describing that they were trying to make it be like her voice in the movie … is pretty bad for them.
It's not precisely copyright, but most states recognize some form of personality rights, which encompass a person's voice just as much as the person's name or visual appearance.
It isn't murky, because law is about intent more than result. It doesn't matter if they hired someone who sounds like Scarlett, it matters if they intended to do so.
If they accidentally hired someone who sounds identical, that's not illegal. But if they intended to, even if it is a pretty poor imitation, it would be illegal because the intent to do it was there.
A court of law would be looking for things like emails about what sort of actress they were looking for, how they described that requirement, how they evaluated the candidate and selected her, and of course, how the CEO announced it alongside a movie title Scarlett starred in.
Under what legal theory is intending to do something which is legal (hiring a person that has a voice you want) becomes illegal because there is another person who has a similar voice?
It's not intending to do something legal, it's intending to do something illegal: Stealing their likeness. The fact you used an otherwise legal procedure to do the illegal activity doesn't make it less illegal.
How can something be illegal if every step towards the objective is legal?
This would result in an incoherent legal system where selective prosecution/corruption is trivial.
It is legal to buy a gun, and legal to fire a gun, and it can even be legal to fire a gun at someone who is threatening to kill you in the moment, but if you fire a gun at someone with the intention of killing someone that happens to be very, very illegal.
So? You're merely (correctly) pointing out that the acts have consequences that are of wildly differing severity. Not that one is a legal and the other is not.
What's illegal, in general, is not the action itself but the intent to do an action and the steps taken in furtherance of that intent.
Hiring someone with a voice you want isn't illegal; hiring someone with a voice you want because it is similar to a voice that someone expressly denied you permission to use is illegal.
Actually, it's so foundational to the common law legal system that there's a specialized Latin term to represent the concept: mens rea (literally 'guilty mind').
Tom Waits successfully sued Frito Lay for using an imitator without approval in a radio commercial.
The key seems to be that if someone is famous and their voice is distinctly attributeable to them, there is a case. In both of these cases, the artists in question were also solicited first and refused.
I feel like that's a little different. In the cases of Midler, Waits, and Johansson, the companies involved wanted to use their voices, were turned down, and then went with an imitator to make it seem to the audience that the celebrity was actually performing. In the case of this "Morgan Freeman" video, Freeman himself is very obviously not performing: the imitator appears on screen, so it's explicitly acknowledged in the ad.
But I'm not a lawyer of any sort either, so... ::shrug::
The Midler v. Ford decision said her voice was distinctive. Not the song.
OpenAI didn't just use a voice like Scarlett Johansson's. They used it in an AI system they wanted people to associate with AI from movies and the movie where Johansson played an AI particularly.[1][2]
You would have to argue the distinctiveness of the voice (if they hadn’t already pursued her to do it). Tom Waits…that’s pretty distinct voice. Scarlett Johansson…not so much
The Tom Waits case had a payout of 2.6 million for services with fair market cost of 100k. What would it cost openai to train chatgpt using her voice? Is she also going to get a payout 26 times that? That GPU budget is starting to look inexpensive...
I’m not a lawyer and don’t have any deep background this area of IP, but there is at least some precedent apparently:
> In a novel case of voice theft, a Los Angeles federal court jury Tuesday awarded gravel-throated recording artist Tom Waits $2.475 million in damages from Frito-Lay Inc. and its advertising agency.
> The U.S. District Court jury found that the corn chip giant unlawfully appropriated Waits’ distinctive voice, tarring his reputation by employing an impersonator to record a radio ad for a new brand of spicy Doritos corn chips.
Yes, absolutely illegal. You don't need to copyright anything, you simply own the rights your own likeness -- your visual appearance and your voice.
A company can't take a photo from your Facebook and plaster it across an advertisement for their product without you giving them the rights to do that.
And if you're a known public figure, this includes lookalikes and soundalikes as well. You can't hire a ScarJo impersonator that people will think is ScarJo.
This is clearly a ScarJo soundalike. It doesn't matter whether it's an AI voice or clone or if they hired someone to sound just like her. Because she's a known public figure, that's illegal if she hasn't given them the rights.
(However, if you generate a synthetic voice that just happens to sound exactly like a random Joe Schmo, it's allowed because Joe Schmo isn't a public figure, so there's no value in the association.)
If you just want ScarJo's (or James Earl Jones') voice, you need the rights from them. Period.
If you want to reuse the character of her AI bot from the movie (her name, overall personality, tone, rhythm, catchphrases, etc.), or the character of Darth Vader, you also need to license that from the producers.
And also from ScarJo/Jones if you want the same voice to accompany the character. (Unless they've sold all rights for future re-use to the producers, which won't usually be the case, because they want to be paid for sequels.)
If they didn’t use her actual voice for the training, didn’t hire voice talent to imitate her, didnt pursue her for a voice contract, didn’t make a reference to the movie in which she voices an AI, I feel OpenAI would have been on more stable legal footing. But they aren’t playing with a strong hand now and folded fast.
Not only that but they didn't credit the voice actress who sounds like her. If she was semi-famous and just naturally sounded like Scarlett Johansson, maybe they could have an argument: "it is not Scarlett, it is the famous [C-list actress] who worked in [production some people may know]".
It’s just that her voice by itself is relatively unremarkable. Someone like say, Morgan freeman, or Barack Obama, someone with a distinctive vocal delivery, that’s one thing. Scarlett Johansson, I couldn’t place her voice out of a lineup. I’m sure it’s pleasant I just can’t think of it.
Scarlett Johansson does absolutely have a distinctive and very famous voice. I wouldn’t take your own ignorance (not meant disparagingly) as evidence otherwise.
That’s why she was the voice actor for the AI voice in Her.
>That’s why she was the voice actor for the AI voice in Her.
She was used in Her because she has a dry/monotone/lifeless form of diction that at the time seemed like a decent stand-in for an non-human AI.
IMDB is riddled with complaints about his vocal-style/diction/dead-pan on every one of her movies. Ghost World, Ghost in the Shell, Lost in Translation, Comic-Book-Movie-1-100 -- take a line from one movie and dub it across the character of another and most people would be fooled, that's impressive given the breadth of quality/style/age across the movies.
When she was first on the scene I thought it was bad acting, but then it continued -- now I tend to think that it's an effort to cultivate a character personality similar to Steven Wright or Tom Waits; the fact that she's now litigating towards protection of her character and likeness reinforces that fact for me.
You know I took some time to compare versus just reading the analysis and in particular I listened to the OpenAI demo and a scene from “her”.
Yeah not moving from my position at all. Just a very generic featureless female voice. I suppose I hear some similarities in timbre, but it’s such an unremarkable voice and diction that it’s hard to put your finger on anything past “generic low affect American alto”.
It’s a great computer voice. Taking it down is for sure the right call PR wise, regardless of whether they may have done.
But given how unscrupulous Sam Altman appears to be, I wouldn't be surprised if OpenAI hired an impersonator as some kind half-ass legal cover, and went about using Johansson's voice anyway. Tech people do stupid shut sometimes because they assume they're so much cleverer than everyone else.
The problem is they pursued, was rejected, then approximated. Had they just approximated and made no references to the movie…then I bet social marketing would have made the connection organically and neither Ms Johansson or the Her producers would have much ground because they could reasonably claim that it was just a relatively generic woman’s voice with a faint NY/NJ accent.
This is known as personality rights or right to publicity. Impersonating someone famous (eg faking their likeness or voice for an ad) is often illegal.
Not here to weigh in on the answers to these questions. But it certainly feels pretty scary to have to ask such questions about a company leading the LLM space, considering the U.S. currently has little to no legal infrastructure to reign in these companies.
Plus the tone of the voice is likely an unimportant detail to theor success. So pushing up against the legal boundaries in this specific domain is at best strange and at worst a huge red flag for their ethics and how they operate.
This is so pointless and petty too. Like "hee hee our software is just like the movies". And continuing the trend of tech moguls watching bleak satire and thinking it's aspirational.
I couldn’t take more than 10s of that. It’s so cringey and for some reason trying to be sexy? Like why would I want a bot to flirt with me while I’m trying to get information?
Seems like sama may have put a big hole in that argument when he tweeted "her", now it is very easy to say that they knowingly cloned ScarJo's likeness. When will tech leaders learn to stop tweeting.
only when they can get a bigger fix from something else.
it takes more than money to fuel these types, and they would have far better minders and bumpers if the downside outweighed the upside. they aren’t stupid, just addicted.
musk was addict smart, owned up to his proclivities and bought the cartel.
Yes, that would be a copyright violation on top of everything else.
Great idea though!
I'm going to start selling Keanu Reeves T-Shirts using this little trick.
See, I'm not using Keanu's likeness if I don't label it as Keanu. I'm just going to write Neo in a Tweet, and then say I'm just cloning Neo's likeness.
Neo is not a real person, so Keanu can't sue me! Bwahahaha
>If you find a guy that looks identical to him, however ...
...it wouldn't make any difference.
A Barack Obama figurine is a Barack Obama figurine, no matter how much you say that it's actually a figurine of Boback O'Rama, a random person that coincidentally looks identically to the former US President.
Any person who knowingly uses another’s name, voice, signature, photograph, or likeness, in any manner, on or in products, merchandise, or goods, or for purposes of advertising or selling, or soliciting purchases of, products, merchandise, goods or services, without such person’s prior consent, or, in the case of a minor, the prior consent of his parent or legal guardian, shall be liable for any damages sustained by the person or persons injured as a result thereof.
I love these takes that constantly pop up in tech circles.
There's no way "you" (the people that engage in these tactics) believe anyone is that gullible to not see what's happening. You either believe yourselves to be exceedingly clever or everyone else has the intelligence of toddler.
With the gumption some tech "leaders" display, maybe both.
If you have to say "technically it's not" 5x in a row to justify a position in a social context just short-circuit your brain and go do something else.
Just to be clear, my comment was sarcastic, I agree with you and I don't think it's a great idea at all.
Writing this comment mostly to say - damn, I didn't think about it this way, but I guess "either believe yourselves to be exceedingly clever or everyone else has the intelligence of toddler" is indeed the mindset.
The only other alternative I can think of is "we all know it's BS, but do they have more money than us to spend on lawyers to call it out?" - which isn't much better TBH.
That’s a weaksauce argument IMO. The character was played by SJ. Depictions of this character necessarily have to be depictions of the voice actress.
Your argument may be stronger if OpenAI said something like “the movie studio owns the rights to this character’s likeness, so we approached them,” but it’s not clear they attempted that.
I honestly don't think Scarlett (the person, not her "her" character) has anything to favor their case, aside from the public's sympathy.
She may have something only if it turns out that the training set for that voice is composed of some recordings of her (the person, not the movie), which I highly doubt and is, unfortunately, extremely hard to prove. Even that wouldn't be much, though, as it could be ruled a derivative work, or something akin to any celebrity impersonator. Those guys can even advertise themselves using the actual name of the celebrities involved and it's allowed.
Me personally, I hope she takes them to court anyway, as it will be an interesting trial to follow.
An interesting facet is, copyright law goes to the substance of the copyrighted work; in this case, because of the peculiarities of her character in "her", she is pretty much only voice, I wonder if that make things look different to the eyes of a judge.
I heard the voice before hearing this news and didn't recognize her, but it's crazy if they really cloned her voice without her permission. Even worse somehow since they did such a bad job at it.
They can say they had a certain thing in mind, which was to produce something like ‘Her’, and obviously Scarjo would have sold it home for them if she participated. But in lieu of the fact that she didn’t, they still went out and created what they had in mind, which was something LIKE ‘Her’. That doesn’t sound illegal.
Not illegal, because this would be a civil case. But they’re on thin ice because there’s a big difference between “creating something like Her” and “creating something like Scarlet Johansson’s performance in Her”.
Creating something like Her is creating something like Scarlet Johansson’s performance of her. The whole point is to hit the same note, which is a voice and an aesthetic and a sensibility. That’s the point! She wasn’t willing to do it. If they hit the same note without training on her voice, then I think that’s fair game.
Yeah, influential people shouldn't get to functionally own a "likeness." It's not a fingerprint. An actor shouldn't credibly worry about getting work because a rich/famous doppelganger exists (which may threaten clientele).
Explicit brand reference? Bad. Circumstantial insinuation? Let it go.
What else does a famous actor have but a likeness? What if their likeness is appropriated to support a cause they disagree with, or a job they declined for moral or ethical reasons? The public will still associate them first and they will bear the consequences.
A famous actor still has acting, name recognition, and generally mountains of cash. But should not get to call dibs on some vague "likeness" simply because they got hot first. Someone shouldn't damage/lose their livelihood because their famous doppelganger iced them out.
I mean, unless an investigation can find any criteria used to select this particular actress like "sounds like Scarlett" in an email somewhere, or you know, the head idiot intentionally and publicly posting the title of a movie starring the actress in relation to the soundalike's voice work.
"when voice is sufficient indicia of a celebrity's identity, the right of publicity protects against its imitation for commercial purposes without the celebrity's consent."
Because it's meant to give the _appearance_ or _perception_ that a celebrity is involved. Their actions demonstrate they were both highly interested and had the expectation that the partnership was going to work out, with the express purpose of using the celebrity's identity for their own commercial purposes.
If they had just screened a bunch of voice actors and chosen the same one no one would care (legally or otherwise).
> Sounds like one of those situations you'd have to prove intent.
The discovery process may help figuring the intent - especially any internal communication before and after the two(!) failed attempts to get her sign-off, as well as any notes shared with the people responsible for casting.
Not necessarily, because this would be a civil matter, the burden of proof is a preponderance of the evidence - it’s glaring obvious that this voice is emulating the movie Her and I suspect it wouldn’t be hard to convince a jury.
What OpenAI did here is beyond the pale. This is open and shut for me based off of the actions surrounding the voice training.
I think a lot of people are wondering about a situation (which clearly doesn’t apply here) in which someone was falsely accused of impersonation based on an accidental similarity. I have more sympathy for that.
But that’s giving OpenAI far more than just the benefit of the doubt: there is no doubt in this case.
I am guessing it's because you are trying to sell the voice as "that" actor voice. I guess if the other voice become popular on its own right (a celebrity) then there is a case to be made.
Whether or not what they've done is currently technically illegal, they're priming the public to be pissed off at them. Making enemies of beloved figures from the broader culture is likely to not make OpenAI many friends.
OpenAI has gone the "it's easier to ask forgiveness than permission" route, and it seemed like they might get away with that, but if this results in a lot more stories like this they'll risk running afoul of public opinion and future legislation turning sharply against them.
Problem is, they asked for permission, showing their hand multiple times. Can I say, I'm somewhat relieved to learn that Sam Altman isn't an evil genius.
I honestly didn't think it sounded like Johansson. Because of the controversy I just now re listened to the demos and I still find if very unlikely that someone would think it was her.
> Though the most successful founders are usually good people, they tend to have a piratical gleam in their eye. They're not Goody Two-Shoes type good. Morally, they care about getting the big questions right, but not about observing proprieties. That's why I'd use the word naughty rather than evil. They delight in breaking rules, but not rules that matter. This quality may be redundant though; it may be implied by imagination.
> Sam Altman of Loopt is one of the most successful alumni, so we asked him what question we could put on the Y Combinator application that would help us discover more people like him. He said to ask about a time when they'd hacked something to their advantage—hacked in the sense of beating the system, not breaking into computers. It has become one of the questions we pay most attention to when judging applications.
it seems most of the big companies try to break the rules while in the process become so strong they trade it off for what becomes a marginal fine & cost of doing business. Facebook, Uber come to mind first. This may just be the same.
What is endearing and admirable as the underdog very easily becomes contemptible abuse in the biggest dog of the pack. It's not a contradiction - the hacking spirit is a trait that causes more damage the more power you have.
The persistent belief in the tech industry as some kind of underdog I think explains much of the recent deplorable behavior by some of the richest people on the planet. A bunch of unbelievably wealthy nerds are mentally trapped in the past and too out of touch to realize they have become the bullies.
I think governance is overly restrictive and stifles innovation. For example I love that Uber and Airbnb exist even though they both sort of skirt or acceptably break rules in place that a complete rule follower wouldn’t have violated.
Taking taxis 15 years ago was an absolute scammy shitty experience and it’s only marginally better now thanks to an actual competitive marketplace
It's actually the most rational thing I've heard quoted from him. You have to be willing to open the box to see what's inside to know if you can do it better/cheaper/faster/smaller. There's ways of doing that without breaking laws, or doing something unethical with the what you learn. There's also ways of doing it without destroying something or violating anyone/anything. It also allows you to hear their response to see if where person is in that rationale. Do they toe the lines, do they run right across it, do they bend but not break, do they scorched earth everything they touch?
Makes me glad there aren't people betting on how far I'm willing to go to bend the law. When you lay it all out like that you make PG sound like a cockfighter paying to get his champion bloodied.
Those are the same people who break the laws and exploit people if they know they can't fight back.
Usually those people are considered sociopaths.
Maybe it's time to ask the employees of OpenAI who fought to get Altman back, How this behavior is compatible with their moral standards or whether money is the most important thing.
Sure, a bit of rebellion can fuel innovation in founders, but as they gain power, it's important to keep things ethical. What seems charming at the startup phase might raise eyebrows as the company expands.
> They delight in breaking rules, but not rules that matter.
To them*
Which is the whole problem. These narcissistic egotists think they, alone, individually, are capable of deciding what's best not just for their companies but for humanity writ large.
> They delight in breaking rules, but not rules that matter.
The question becomes "what rules matter?". And the answer inevitably becomes "only the ones that work in my favor and/or that I agree with".
I think someone trying to defend this would go "oh come on, does it really matter if a rich actress gets slightly richer?" And no, honestly, it doesn't matter that much. Not to me, anyway. But it matters that it establishes (or rather, confirms and reinforces) a culture of disregard and makes it about what you think matters, and not about what someone else might think matters about the things in their life. Their life belongs to them, a fact that utopians have forgotten again and again everywhere and everywhen. And once all judgement is up to you, if you're a sufficiently ambitious and motivated reasoner (and the kind of person we're talking about here is), you can justify pretty much whatever you want without that pesky real-world check of a person going "um actually no I don't want you to do that".
Sometimes I think anti-tech takes get this wrong. They see the problem as breaking the rules at all, as disrupting the status quo at all, as taking any action that might reasonably be foreseen to cause harm. But you do really have to do that if you want to make something good sometimes. You can't foresee every consequence of your actions - I doubt, for example, that Airbnb's founders were thinking about issues with housing policy when they started their company. But what differentiates behavior like this from risk-taking is that the harm here is deliberate and considered. Mistakes happen, but this was not a mistake. It was a choice to say "this is mine now".
That isn't a high bar to clear. And I think we can demand that tech leaders clear it without stifling the innovation that is tech at its best.
So a kind of lack of empathy?
Do these guys have this image of "autists" and are basically filtering for them, cause this criteria seems to be favoring oppositional defiance disorder.
I mention this specifically because I remember mark andreseen comment something similar in lex fridman's podcast, something along the lines of getting "those creative people" together to build on ai.
It was posted by the tech reporter at NPR. Inb4 “journos can’t be trusted” blah blah blah, here in reality NPR is a reputable org and a reasonable person’s Bayesian priors would put this at “almost certainly an actual statement from ScarJo.”
Variety has a story.[1] It doesn't yet mention an direct statement from Johannson. But watch that space. Variety is well connected in Hollywood and will check with her agent to confirm or deny.
Variety article updated: [UPDATE: Johansson released a statement saying Altman had reached out to ask her to lend her voice to ChatGPT but she declined; when she heard the demo, “I was shocked, angered and in disbelief that Mr. Altman would pursue a voice that sounded so eerily similar to mine.”]
You have no idea. Th will settle or be forced to admit they used a movie studios IP without payment to clone a voice model. They will cut a check for tens of millions and maybe stock as well. They will run from this as is clearly obvious from the immediate takedown. They are in crisis mode.
I don't agree, if Johansson is serious about wanting a legal precedent set and doesn't care about the money (tbd) then a hypothetical lawsuit could go to the US Supreme court and lead to a decision that has significant ramifications.
Many of OpenAI's productization ideas make more sense when you remember that the guy in charge also thought Worldcoin and it's eye scanning or were a good idea.
I think it’s interesting that Johansson chose to forgo substantial royalties and collaboration potential
But it must feel pretty fucking weird and violatory when you spend your entire life thinking about how you are going to deliver certain lines and that’s your creative Body of work, and then for someone to just take that Voice and apply it to any random text that can be generated?
I get why she wouldn’t want to let it go.
In a way it is similar to how a developer might feel about their code being absorbed, generalized, and then regurgitated almost verbatim as part of some AI responses
But in the case of voice it’s even worse as the personality impression is contained in the slightest utterance… whereas a style of coding Or a piece of code might be less Recognizable, and generally applicable to such a wide range of productions
Voice is the original human technology, To try to take that from someone without their consent is a pretty all encompassing grab
That is not her only appeal, she is a renowned actress.
However Sam tweeted "her" which is literally the movie where she voices the AI girlfriend. And then made a synthetic replica of her voice the star of their new demo against her wishes.
I don’t think it’s about me, but since you brought it up, I try to maintain my innocence in this world. I try to be biased towards that rather than cynicism, at least… I think that’s important. Cynicism is a kind of arrogance: where you think you’ve seen it all before… but you’re wrong.
But thank you for the opportunity Your comment provides to speak to that: I do appreciate it. the chance to add clarity.
ScarJo probably has enough money for ten lifetimes already. The potential downside of signing your identity away to a SV plutocrat with questionable morality and world-changing technology is enormous. And it seems like she was instantly vindicated in not trusting him.
Were royalties in the picture? Don't think it's crazy to think that a company obsessed with replacing artists wouldn't value one enough to pay royalties.
And in terms of collaboration potential... OpenAI is a big draw for businesses and a subset of tech enthusiasts, but I don't think artists in any industry are dying to collaborate with them.
SJ is a major character in like 5 of the top grossing movies of all time. The royalties from her voice being used by ChatGPT would be meaningless to her
OpenAI has successfully stolen the intellectual property of millions of people to incorporate into their product, so why would they fear stealing someones voice at this point? I hope she wins. Maybe it'll set some kind of precedent.
OpenAI is trying to demonstrate how it's so trustworthy, and is always talking about how important it is to be trustworthy when it comes to something as important and potentially dangerous as AI.
And then they do something like this...??
I literally don't understand how they could be this dumb. Do they not have a lawyer? Or do they not tell their corporate counsel what they're up to? Or just ignore the counsel when they do?
Especially considering that the advantage gained by having an AI sound like this one individual is absolutely minimal. It's not as though any significant portion of a target market is going to throw a tantrum, saying "No, no, I refuse to accept this simulated companionship unless it sounds exactly like the voice in that one particular movie several years ago." Baffling that the company didn't recognise the risk here and retire that voice as soon as they were turned down the first time.
This demo and the one I linked just seem so open about the AI GF use case its bizarre.
If you actually wanted a voice assistant AI, having a giggly, chatty computer acting like it has a huge crush on you is not remotely useful in day to day real world use. Unless that's exactly what you want.
I can't fathom such a bad decision as asking someone for permission to use their voice and doing it anyway after they say no. It's almost like NYT is currently suing them for unauthorized use and they should really not be making such an amateur mistake.
Not only that, I don’t think Disney wants this precedent. They may also want to get back on her good side. Either way given they own a number of movies with her in starring roles I wouldn’t be surprised if they were happy to help in her lawsuit with the legal fees.
Hell they may sue on their own or join as another damaged party.
I’m sure they are salivating at the thought of discovery forcing OAI to crack open their datasets so they can put a dollar amount to every piece of infringing material in there.
The way USA courts are set up, setting precedent and assessing damages are two distinct things. I agree that the precedent she would be targeting wouldn’t be all that financially rewarding but that’s not the only thing that motivates humans.
The demo OpenAI was a massive marketing campaign for GPT-4o and led to the largest increases in revenue for their mobile app. The voice was a large part of why this release was a hit. The demo is still on youtube with 4M views. She has a great case for financial remuneration even if they haven't yet launched the voice feature.
They didn’t come to an agreement to use her voice, they used her voice. And they obviously knew it was a problem because they went back to her like the night before trying to get her approval again.
The correct thing to do was NOT use her voice.
You don’t get to steal something, get all the benefit from it (the press coverage), and then say “oops never mind it was just a few hours you can’t sue us”.
Why don’t we try selling tickets to watch a Disney movie “just one time” and see how well that goes. I don’t think Disney’s lawyers will look at it and say “oh well they decided not to do it again.“
They make no representation that it is her voice, and there's a really good chance that they separately made a voice that sounds similar enough to where if they could tack her name to it, it'd be good for advertising, but otherwise isn't her voice.
Voices are really hard for people to distinguish as being a certain person without priming, so really she's doing for free the advertising they were hoping she'd do for pay
> there's a really good chance that they separately made a voice that sounds similar enough to where if they could tack her name to it, it'd be good for advertising, but otherwise isn't her voice.
There’s also a really good chance this is in some way a deepfake. Would be interesting to see this get examined by courts.
Not OP, or the replier, but I think the longer answer here is that if Scarlett Johansson <insert any wealthy and/or popular figure> can't win a lawsuit against a company that has effectively:
(1) Used content she has created (vocal lines) in order to train a generative AI with the ability to create more of that content (vocal lines), without permission or payment, let alone acknowledgement
(2) In doing so removed the scarcity of the content she provides (Generative AI is effectively unlimited in the lines it can produce once effectively trained)
Then no smaller time actor, voice actor, artist, musician, etc, is likely to have any chance defending themselves against the theft of their work for AI purposes.
And, with that precedent set, the legal and financial landscape of art and creativity will have changed in a way that discourages anybody from creating original works for financial gain, because we've systematised the creation of original and derivative works with no legal or financial ramifications.
Maybe, but I don't think it's likely that they used any actual copyrighted material of her voice. It's not illegal to want one voice actor to play a role (or record voice data), then get a different, somewhat (but not actually very) similar sounding voice actor to do it instead.
The above comment from MBCook captures it well but basically I would say that for the majority of people I know:
Cons: the rise of scams, carbon emissions, and, some think, existential risk.
Pros: ??? You can be more efficient at the email job you hate. And it gives you advice on gardening and home improvement so you don’t have to call your uncle.
Um, okay. The original question of “curious, why?” didn’t imply that I need respond with any level of rigor. I don’t think that the burden of proof should be on me to explain why ChatGPT is useful when I don’t believe it to be. But please feel free to not bother to pick apart my opinion ”word-by-word”. No one asked you to
Sam Altman has a smarmy, borderline vomit inducing face. The fact that I have to see his face every day while scrolling through the interwebs feeds has made my life worse.
See, I don't like Sam Altman either, but this habit of criticizing people for their faces seems wrong-headed. Isn't it his behavior that we should be criticizing?
I could be wrong but in these cases (or when people criticize voices, similarly), people are more put off by expressions than raw features. Nonverbal communication is high emotional bandwidth. I don’t begrudge someone disliking someone else’s self presentation.
Very obvious bias against OpenAI in the comments here. Possible motives: a) there's a visceral human reaction against anyone extraordinarily successful and powerful (Nietzsche wrote about this). b) resentment for OpenAI's advancements in code generation and its possible impact on the job market. I don't think much of the outrage here is motivated by altruism, it's probably more about siding with whoever opposes your perceived enemy.
There's any number of people with equivalent voices that Scarlet has stolen the voices of by selling it to movies, and all of those other people who sound the same deserve compensation from her for it
Most people aren’t that selfish or petty. But there are some people who are—and by an unfortunate quirk of human nature, they tend to believe everyone else is just like them.
Uh, what? This is a pretty well-researched phenomenon: A very high percentage of people who engage in pathological behaviors that take advantage of other people believe that “everyone does it” and rely on that as justification. Penn & Teller did a pithy bit on the topic.
Can you provide the evidence indicating that her voice was stolen? Maybe also an audio analysis of the vocal characteristics involved and the degree of similarity? And wouldn't it be prudent to investigate these matters before making such accusations?
Never attribute to jealousy that which can be explained by a difference of perspective.
Jealousy can explain any criticism of anybody. That makes it an epistemic hazard; you can always fall back to it, and you can develop a habit of doing so. And then you have deafened yourself. Any substantive criticisms will be lost on you. As will any opportunity to learn from those you disagree with. Your ideas will be hot house flowers, comfortable and safe in their controlled environment, but unable to contend in the wild. Aspire to weeds.
That would be tautological, as 'difference of perspective' is a restatement of the phenomenon at a higher level of abstraction. Someone who is resentful will likely have a different perspective than they would otherwise have. Psychological motives are real and don't disappear by simply assuming a variant of a problem-solving heuristic. Notice also I didn't use the word 'jealousy.' I do think that speculations about psychological motives should be made only after careful consideration and generally remain unprovable. However, the uncertainty of psychological motives shouldn't prevent us from ever addressing them.
You can use whatever high minded references to philosophy and psychology you please, but from what I've read from you in this thread - you're missing the point. You aren't understanding the criticism, and you're addressing the weakest form of the argument. I'm suggesting you aspire to address the strongest form, whether the weakest is present or not.
If you choose to do nothing with that information, that's up to you.
You could go to Penguin directly. Or perhaps the Authors Guild would help you set something up across publishers.
All of deviant art? Probably not. But do you need to train on all of that? You could certainly run ads telling people you’d be willing to pay a small amount to train on their art and let them choose.
Would it be legal to train against the national archives?
Options exist. No you won’t get as much stuff as you do by taking whatever you want, but people are being compensated for their work or at least being given the choice to opt in.
Ignoring the actual legality of using training data in machine learning, let's look at these "options" in a purely objective-driven way. If you do that, you'll quickly realize that what you describe would strengthen megacorporations to an even greater extent. If training on public data is banned, then the entities who have complete ownership of all their data would be granted a de facto monopoly. Stock image services, media conglomerates, music labels, industry giants would start in the winning position. When one side just gets their way for free and the other has to beg for scraps, do you really think this is a fair proposition?
The reason why open-source AI exists at all is that we've always been allowing use of public data - it was okay when Google did it, it was okay when the Internet Archive did it, it was even okay when text translation services used that same data to train their models - or really, that applies to basically anything ML-driven before generative AI.
There's, like, a sea of reasons to criticize OpenAI for - but arguing for extending IP laws even further and calling out opponents for "literal theft" is one of the weaker options that caught on with many people.
Those sound like really large obstacles. Also, projecting this thinking forward, if we have humanoid robots walking around assisting humans, should they not be able to learn from their environment without paying every time they look at a piece of art or read something?
you're not wrong, and i think the technology exists to accomplish something that did this, and has for a few years. but the end result would be a system that funnels money from large numbers of people, to different large numbers of people. and a lot of overlap and cross payments too. maybe there's a good business in transaction fees on that? but it seems like big numbers going to other people while a much smaller number goes to the party owning the system is not an appealing business in our world.
ipso facto why does it not exist while spotify buys podcasts and kills them
Creatives will use AI even more than regular people and with better results. What I find dangerous is to protect all paraphrases and variations as well, akin to owning an idea. When not replicating expression ideas should be free, even in copyrights works
I think majority agrees that we're ok with a few millionaire priviledge actors losing small amount of their value in favor of personal AI assistants for general populace. How is that even a topic worth discussing is trully perplexing.
Why? The data clearly points that nobody will go out of their way to protect IP laws. In majority of the world nobody could care less about some hollywood actor with tough to spell name getting their voice cloned.
It's really weird that someone would think that IP law is more important than access to information. Must be some dystopian bubble.
Just because companies are violating people's artistic ownership and the State isn't punishing or preventing it fast enough is not an indication that a majority of people "could care less". If anything you can see the fomenting dislike with OpenAI right in front of you.
You can frame it however you like to make it justifiable, but I'm not going to get into the mud with you.
You're building a strawman here. The topic is whether information access is more important in our society than IP protection and the answer is big yes no matter how you look at it and I don't see you ever winning this argument. Take care o/
I don’t know a single person who has ever mentioned wanting an AI assistant, and I know a lot of people who are creeped out by the idea. I live in a swing state and work for a big tech company but most of my friends aren’t in tech. So I’m not sure that a “majority agrees” about anything related to an AI assistant
Yet every single person uses an AI assistant every day when they Google something or ask their phone for who performs the song that's playing in the background. You must know some of these people?
I don’t think that I would call every AI feature an AI Assistant, but sure, if that’s how you define “AI Assistant,” then you’re correct that I do in fact know people who use/want one. But I think that Google+AI or a music player is a pretty different product experience/design/market fit than a “personal AI Assistant” like an AI-powered Siri or Alexa or Google Assistant. I don’t know anyone who uses (or wants to use) the latter group for anything other than “Alexa, play XYZ”, or “Siri, set an alarm” which I wouldn’t really call GPT-4-level AI
Just because LLMs only really existed for a year now and it's still the fastest adopted new tech we ever had. Even iphone took longer to adopt. It'll take time for LLMs to lodge into direct use and thus the move to voice chat by openAI but every single person is already full in AI game from photo editing, to content curation - AI assists us everywhere already. The leap between that and texting to an AI friend is extremely tiny.
They stole someone’s voice after being told not to?
They unleashed a massive new wave of spam and scams and garbage onto the Internet?
Because they’re stealing every single bit of content on the Internet and everywhere else they can get their hands on without paying anything for it and then expect to sell it back to us in chewed up garbage form?
They help single-handedly cause MS to blast past their carbon commitments by 30%? And got everyone else into a big race for how many resources they could waste to power AI nonsense that doesn’t even actually work that well for what people want to use it for?
Perhaps the fact that they don’t seem to care one bit about any of the consequences, legal/moral/ethical/economical/etc. caused by what they’ve done as long as they make money?
I don’t have a grudge against open AI. I have a grudge against the AI industry and the way these kind of SV golden boys seem to think they’re immune from criticism.
Why do you think stealing a professional actor’s voice should be OK and immune from criticism? This is horrible.
Amen. We need ethics and accountability and this wave of “AI” has been sorely lacking those, instead preferring the Uber model of “let’s just get big enough to bully things into working out in our favor.”
It’s important to nip that shit in the bud lest it spread.
This is hilarious. OpenAI didn’t even need to press for this voice, their technical demo was impressive enough, but they did and now it’ll cast a shadow over a pretty impressive AI advancement. In the long term though, this won’t matter.
1,065 comments
[ 3.6 ms ] story [ 392 ms ] threadOf course, Twitter continues to bring people with big egos to their own downfall.
Maybe only when the director's instructions are "I want you to sound like XYZ".
>The Ninth Circuit reversed the District Court, finding that White had a cause of action based on the value of her image, and that Samsung had appropriated this image. Samsung's assertion that this was a parody was found to be unavailing, as the intent of the ad was not to make fun of White's characteristics, but to sell VCRs.
https://en.wikipedia.org/wiki/White_v._Samsung_Electronics_A....
Maybe it depends on which court will handle the case, but OpenAI's core intent isn't parody, but rather to use someone's likeness as a way to make money.
(I am not a lawyer)
It would be harder to find a case if they simply just hired someone that sounds similar, but if they did that with the intention to sound like the original that's still impersonation, only it's harder to prove.
If they just happened to hire someone that sounded like the original, then that's fair game IMO.
IANAL
mainstream adoption hasn't been that great - now there's drama
[0]: https://www.mercurynews.com/2021/07/20/how-the-doodle-god-un...
[1]: https://youtu.be/-CopbQ_QgmM?si=gkbWEva_qqG8dTib&t=205
It is a silly place.
It's not just successful companies though. There is a bit of ego necessary in a founder that makes them think their idea or their implementation of a thing is better so that it needs to be its own company. Sometimes though they even get caught up in their own reality distortion fields with obviously bad ideas or ideas implemented badly due to their own arrogance that ultimately fails.
Open for business, Open to suggestions, and Open season for any lawyers that want a piece of the sizable damages.
If I thought I had a great idea, I would want people to try to poke holes in it. Yet founders often universally seem to be the incredibly sensitive and insecure about their idea.
[1] https://www.lesswrong.com/posts/QDczBduZorG4dxZiW/sam-altman...
[2] https://www.hackingbutlegal.com/p/statement-by-annie-altman-...
That is...not a pretty picture. We desperately need someone else at the helm of OpenAI.
If they accidentally hired someone who sounds identical, that's not illegal. But if they intended to, even if it is a pretty poor imitation, it would be illegal because the intent to do it was there.
A court of law would be looking for things like emails about what sort of actress they were looking for, how they described that requirement, how they evaluated the candidate and selected her, and of course, how the CEO announced it alongside a movie title Scarlett starred in.
In the case of acquiring a likeness, if it's done legally you acquire someone else's likeness that happens to be shared with your target.
The likeness is shared and non-unique.
If you objective is to take someone's life, there is no other pathway to the objective but their life. With likeness that isn't the case.
Hiring someone with a voice you want isn't illegal; hiring someone with a voice you want because it is similar to a voice that someone expressly denied you permission to use is illegal.
Actually, it's so foundational to the common law legal system that there's a specialized Latin term to represent the concept: mens rea (literally 'guilty mind').
Yes, it will be interesting in June 1988 when we will find out "where this lands": https://en.wikipedia.org/wiki/Midler_v._Ford_Motor_Co.
https://en.m.wikipedia.org/wiki/Midler_v._Ford_Motor_Co.
Bette Middler successfully sued Ford for impersonating her likeness in a commercial.
Then also:
https://casetext.com/case/waits-v-frito-lay-inc
Tom Waits successfully sued Frito Lay for using an imitator without approval in a radio commercial.
The key seems to be that if someone is famous and their voice is distinctly attributeable to them, there is a case. In both of these cases, the artists in question were also solicited first and refused.
https://www.hollywoodreporter.com/business/business-news/bac...
But I'm not a lawyer of any sort either, so... ::shrug::
OpenAI didn't just use a voice like Scarlett Johansson's. They used it in an AI system they wanted people to associate with AI from movies and the movie where Johansson played an AI particularly.[1][2]
[1] https://blog.samaltman.com/gpt-4o
[2] https://x.com/sama/status/1790075827666796666
> In a novel case of voice theft, a Los Angeles federal court jury Tuesday awarded gravel-throated recording artist Tom Waits $2.475 million in damages from Frito-Lay Inc. and its advertising agency.
> The U.S. District Court jury found that the corn chip giant unlawfully appropriated Waits’ distinctive voice, tarring his reputation by employing an impersonator to record a radio ad for a new brand of spicy Doritos corn chips.
https://www.latimes.com/archives/la-xpm-1990-05-09-me-238-st...
A company can't take a photo from your Facebook and plaster it across an advertisement for their product without you giving them the rights to do that.
And if you're a known public figure, this includes lookalikes and soundalikes as well. You can't hire a ScarJo impersonator that people will think is ScarJo.
This is clearly a ScarJo soundalike. It doesn't matter whether it's an AI voice or clone or if they hired someone to sound just like her. Because she's a known public figure, that's illegal if she hasn't given them the rights.
(However, if you generate a synthetic voice that just happens to sound exactly like a random Joe Schmo, it's allowed because Joe Schmo isn't a public figure, so there's no value in the association.)
If you imitate Darth Vader, I don't think James Earl Jones has as much case for likeliness as Star Wars franchise
If you just want ScarJo's (or James Earl Jones') voice, you need the rights from them. Period.
If you want to reuse the character of her AI bot from the movie (her name, overall personality, tone, rhythm, catchphrases, etc.), or the character of Darth Vader, you also need to license that from the producers.
And also from ScarJo/Jones if you want the same voice to accompany the character. (Unless they've sold all rights for future re-use to the producers, which won't usually be the case, because they want to be paid for sequels.)
https://en.m.wikipedia.org/wiki/Midler_v._Ford_Motor_Co.
Just find someone who sounds like her, then hire them for the rights to their voice.
There is no doubt that the hired actor was an impersonator, this was explicitly stated by scama himself.
That’s why she was the voice actor for the AI voice in Her.
She was used in Her because she has a dry/monotone/lifeless form of diction that at the time seemed like a decent stand-in for an non-human AI.
IMDB is riddled with complaints about his vocal-style/diction/dead-pan on every one of her movies. Ghost World, Ghost in the Shell, Lost in Translation, Comic-Book-Movie-1-100 -- take a line from one movie and dub it across the character of another and most people would be fooled, that's impressive given the breadth of quality/style/age across the movies.
When she was first on the scene I thought it was bad acting, but then it continued -- now I tend to think that it's an effort to cultivate a character personality similar to Steven Wright or Tom Waits; the fact that she's now litigating towards protection of her character and likeness reinforces that fact for me.
It's unique to her though , that's for sure.
Do you have a source for this?
Yeah not moving from my position at all. Just a very generic featureless female voice. I suppose I hear some similarities in timbre, but it’s such an unremarkable voice and diction that it’s hard to put your finger on anything past “generic low affect American alto”.
It’s a great computer voice. Taking it down is for sure the right call PR wise, regardless of whether they may have done.
And here's some caselaw where another major corporation got smacked down for doing the exact same thing: https://en.wikipedia.org/wiki/Midler_v._Ford_Motor_Co.
But given how unscrupulous Sam Altman appears to be, I wouldn't be surprised if OpenAI hired an impersonator as some kind half-ass legal cover, and went about using Johansson's voice anyway. Tech people do stupid shut sometimes because they assume they're so much cleverer than everyone else.
https://en.wikipedia.org/wiki/Personality_rights
Plus the tone of the voice is likely an unimportant detail to theor success. So pushing up against the legal boundaries in this specific domain is at best strange and at worst a huge red flag for their ethics and how they operate.
And not see how over the top it is... cmon.
Tech Company: At long last, we have created the Torment Nexus from classic sci-fi novel Don't Create The Torment Nexus
https://x.com/AlexBlechman/status/1457842724128833538?lang=e...
If so, I suspect they’ll be okay in a court of law — having a voice similar to a celebrity isn’t illegal.
It’ll likely cheese off actors and performers though.
[1] https://www.forbes.com/sites/roberthart/2024/05/20/openai-sa...
it takes more than money to fuel these types, and they would have far better minders and bumpers if the downside outweighed the upside. they aren’t stupid, just addicted.
musk was addict smart, owned up to his proclivities and bought the cartel.
Is there a distinction?
Are they trying to make it sound like Her, or SJ? Or just trying to go for a similar style? i.e. making artistic choices in designing their product
Note: I've never watched the movie.
Yes, that would be a copyright violation on top of everything else.
Great idea though!
I'm going to start selling Keanu Reeves T-Shirts using this little trick.
See, I'm not using Keanu's likeness if I don't label it as Keanu. I'm just going to write Neo in a Tweet, and then say I'm just cloning Neo's likeness.
Neo is not a real person, so Keanu can't sue me! Bwahahaha
...it wouldn't make any difference.
A Barack Obama figurine is a Barack Obama figurine, no matter how much you say that it's actually a figurine of Boback O'Rama, a random person that coincidentally looks identically to the former US President.
That situation would be completely legal. Come on.
California Civil Code Section 3344(a) states:
Any person who knowingly uses another’s name, voice, signature, photograph, or likeness, in any manner, on or in products, merchandise, or goods, or for purposes of advertising or selling, or soliciting purchases of, products, merchandise, goods or services, without such person’s prior consent, or, in the case of a minor, the prior consent of his parent or legal guardian, shall be liable for any damages sustained by the person or persons injured as a result thereof.
Not make a living posing for pictures without consent of the said celebrity?
Re: "get real" - the law is pretty real.
Things like parody are protected under fair use, explicitly.
There's no way "you" (the people that engage in these tactics) believe anyone is that gullible to not see what's happening. You either believe yourselves to be exceedingly clever or everyone else has the intelligence of toddler.
With the gumption some tech "leaders" display, maybe both.
If you have to say "technically it's not" 5x in a row to justify a position in a social context just short-circuit your brain and go do something else.
Writing this comment mostly to say - damn, I didn't think about it this way, but I guess "either believe yourselves to be exceedingly clever or everyone else has the intelligence of toddler" is indeed the mindset.
The only other alternative I can think of is "we all know it's BS, but do they have more money than us to spend on lawyers to call it out?" - which isn't much better TBH.
Your argument may be stronger if OpenAI said something like “the movie studio owns the rights to this character’s likeness, so we approached them,” but it’s not clear they attempted that.
As to whether she owns the rights of that performance or somebody else, we'd have to read the contract; most likely she doesn't, though.
She may have something only if it turns out that the training set for that voice is composed of some recordings of her (the person, not the movie), which I highly doubt and is, unfortunately, extremely hard to prove. Even that wouldn't be much, though, as it could be ruled a derivative work, or something akin to any celebrity impersonator. Those guys can even advertise themselves using the actual name of the celebrities involved and it's allowed.
Me personally, I hope she takes them to court anyway, as it will be an interesting trial to follow.
An interesting facet is, copyright law goes to the substance of the copyrighted work; in this case, because of the peculiarities of her character in "her", she is pretty much only voice, I wonder if that make things look different to the eyes of a judge.
That falls under copyright, trademarks, ...
People who want to use an actor's likeness can't get around likeness rights by saying they impersonated a specific performance actually.
Explicit brand reference? Bad. Circumstantial insinuation? Let it go.
"when voice is sufficient indicia of a celebrity's identity, the right of publicity protects against its imitation for commercial purposes without the celebrity's consent."
If they had just screened a bunch of voice actors and chosen the same one no one would care (legally or otherwise).
(and given the timeline ScarJo laid out in her Twitter feed, I'd be inclined to vote to convict at the present moment)
The discovery process may help figuring the intent - especially any internal communication before and after the two(!) failed attempts to get her sign-off, as well as any notes shared with the people responsible for casting.
I think a lot of people are wondering about a situation (which clearly doesn’t apply here) in which someone was falsely accused of impersonation based on an accidental similarity. I have more sympathy for that.
But that’s giving OpenAI far more than just the benefit of the doubt: there is no doubt in this case.
This isn't actually complicated at all. OpenAI robbed her likeness against her express will.
There's not a lot of precedent around voice impersonation, but there is for a very, very similar case against Ford
https://en.m.wikipedia.org/wiki/Midler_v._Ford_Motor_Co.
https://www.quimbee.com/cases/waits-v-frito-lay-inc
https://en.wikipedia.org/wiki/White_v._Samsung_Electronics_A....
OpenAI has gone the "it's easier to ask forgiveness than permission" route, and it seemed like they might get away with that, but if this results in a lot more stories like this they'll risk running afoul of public opinion and future legislation turning sharply against them.
> Though the most successful founders are usually good people, they tend to have a piratical gleam in their eye. They're not Goody Two-Shoes type good. Morally, they care about getting the big questions right, but not about observing proprieties. That's why I'd use the word naughty rather than evil. They delight in breaking rules, but not rules that matter. This quality may be redundant though; it may be implied by imagination.
> Sam Altman of Loopt is one of the most successful alumni, so we asked him what question we could put on the Y Combinator application that would help us discover more people like him. He said to ask about a time when they'd hacked something to their advantage—hacked in the sense of beating the system, not breaking into computers. It has become one of the questions we pay most attention to when judging applications.
"What We Look for in Founders", PG
https://paulgraham.com/founders.html
I think the more powerful you become, the less endearing this trait is.
(Which wasn't even the taxi drivers, although they were plenty bad enough on their own.)
Taking taxis 15 years ago was an absolute scammy shitty experience and it’s only marginally better now thanks to an actual competitive marketplace
Usually those people are considered sociopaths.
Maybe it's time to ask the employees of OpenAI who fought to get Altman back, How this behavior is compatible with their moral standards or whether money is the most important thing.
To them*
Which is the whole problem. These narcissistic egotists think they, alone, individually, are capable of deciding what's best not just for their companies but for humanity writ large.
> They delight in breaking rules, but not rules that matter.
The question becomes "what rules matter?". And the answer inevitably becomes "only the ones that work in my favor and/or that I agree with".
I think someone trying to defend this would go "oh come on, does it really matter if a rich actress gets slightly richer?" And no, honestly, it doesn't matter that much. Not to me, anyway. But it matters that it establishes (or rather, confirms and reinforces) a culture of disregard and makes it about what you think matters, and not about what someone else might think matters about the things in their life. Their life belongs to them, a fact that utopians have forgotten again and again everywhere and everywhen. And once all judgement is up to you, if you're a sufficiently ambitious and motivated reasoner (and the kind of person we're talking about here is), you can justify pretty much whatever you want without that pesky real-world check of a person going "um actually no I don't want you to do that".
Sometimes I think anti-tech takes get this wrong. They see the problem as breaking the rules at all, as disrupting the status quo at all, as taking any action that might reasonably be foreseen to cause harm. But you do really have to do that if you want to make something good sometimes. You can't foresee every consequence of your actions - I doubt, for example, that Airbnb's founders were thinking about issues with housing policy when they started their company. But what differentiates behavior like this from risk-taking is that the harm here is deliberate and considered. Mistakes happen, but this was not a mistake. It was a choice to say "this is mine now".
That isn't a high bar to clear. And I think we can demand that tech leaders clear it without stifling the innovation that is tech at its best.
I mention this specifically because I remember mark andreseen comment something similar in lex fridman's podcast, something along the lines of getting "those creative people" together to build on ai.
[1] https://variety.com/2024/digital/news/openai-pulls-scarlett-...
Stuff from her comes via press agents, which is generally sent directly to reporters.
https://nitter.poast.org/BobbyAllyn/status/17926794357010149...
Yeah. _That_ well known completely rational and unquestionably incorruptible institution.
I would bet Altman had been to more teenage sex parties and paid for more holidays with SC judges that Scarlett has... :sigh:
But it must feel pretty fucking weird and violatory when you spend your entire life thinking about how you are going to deliver certain lines and that’s your creative Body of work, and then for someone to just take that Voice and apply it to any random text that can be generated?
I get why she wouldn’t want to let it go.
In a way it is similar to how a developer might feel about their code being absorbed, generalized, and then regurgitated almost verbatim as part of some AI responses
But in the case of voice it’s even worse as the personality impression is contained in the slightest utterance… whereas a style of coding Or a piece of code might be less Recognizable, and generally applicable to such a wide range of productions
Voice is the original human technology, To try to take that from someone without their consent is a pretty all encompassing grab
Not a bad call for someone already rich
To suggest that Johansson’s only appeal is to the opposite gender (and ‘lonely’ ones at that!) I think is myopic and reductive of her impact
However Sam tweeted "her" which is literally the movie where she voices the AI girlfriend. And then made a synthetic replica of her voice the star of their new demo against her wishes.
It's pretty direct what he is pitching at.
but it doesn’t matter, because how he may have marketed it in a 140 character tweet does not encompass the entirety of how it could be used, of course
So what? Reducing a situation to its essentials is an entirely valid argument / debating technique.
lol hahaha!
That said, Krazam covered this topic well already https://www.youtube.com/watch?v=KiPQdVC5RHU
Also, suddenly your voice is being used to say things that you would never say. What could go wrong?
Suddenly, instead of ScarJo, the famous movie star, you are that crazy voice from OpenAI.
To suggest that Johansson’s only appeal is to the opposite gender (and ‘lonely’ ones at that!) I think is myopic and reductive of her impact
hahaha! :)
I don’t think it’s about me, but since you brought it up, I try to maintain my innocence in this world. I try to be biased towards that rather than cynicism, at least… I think that’s important. Cynicism is a kind of arrogance: where you think you’ve seen it all before… but you’re wrong.
But thank you for the opportunity Your comment provides to speak to that: I do appreciate it. the chance to add clarity.
And in terms of collaboration potential... OpenAI is a big draw for businesses and a subset of tech enthusiasts, but I don't think artists in any industry are dying to collaborate with them.
"Each actor receives compensation above top-of-market rates, and this will continue for as long as their voices are used in our products."
https://openai.com/index/how-the-voices-for-chatgpt-were-cho...
Not sure if this is royalties, but it seems like there's some form of long term compensation. But it's a little vague so not sure.
We know nothing about their offer to her. Could have just been a bad deal
The wink wink at creating an AI girlfriend is so bizarre
I guess we know who their target user base is
OpenAI is trying to demonstrate how it's so trustworthy, and is always talking about how important it is to be trustworthy when it comes to something as important and potentially dangerous as AI.
And then they do something like this...??
I literally don't understand how they could be this dumb. Do they not have a lawyer? Or do they not tell their corporate counsel what they're up to? Or just ignore the counsel when they do?
He’s absolutely been the bad guy all along.
[1] https://www.youtube.com/watch?v=0AjqljwVusk
https://twitter.com/OpenAI/status/1790089521985466587
Giggly, flirty AI voice demos were already weird, but now it's even creepier knowing the backstory of how they try to get their voices.
If you actually wanted a voice assistant AI, having a giggly, chatty computer acting like it has a huge crush on you is not remotely useful in day to day real world use. Unless that's exactly what you want.
She has the resources to fight back and make an example of them, and they have the resources to make it worthwhile.
Hell they may sue on their own or join as another damaged party.
They didn’t come to an agreement to use her voice, they used her voice. And they obviously knew it was a problem because they went back to her like the night before trying to get her approval again.
The correct thing to do was NOT use her voice.
You don’t get to steal something, get all the benefit from it (the press coverage), and then say “oops never mind it was just a few hours you can’t sue us”.
Why don’t we try selling tickets to watch a Disney movie “just one time” and see how well that goes. I don’t think Disney’s lawyers will look at it and say “oh well they decided not to do it again.“
Voices are really hard for people to distinguish as being a certain person without priming, so really she's doing for free the advertising they were hoping she'd do for pay
There’s also a really good chance this is in some way a deepfake. Would be interesting to see this get examined by courts.
But if they concocted a fake voice to sound as much like her as possible, that’s not really better.
Altman’s tweet, combined with previous statements Her is his favorite movie, and trying to secure rights twice looks really really damning.
> so really she's doing for free the advertising they were hoping she'd do for pay
They didn’t want her to advertise for them. They wanted to use her voice. Do you not see a difference?
(1) Used content she has created (vocal lines) in order to train a generative AI with the ability to create more of that content (vocal lines), without permission or payment, let alone acknowledgement
(2) In doing so removed the scarcity of the content she provides (Generative AI is effectively unlimited in the lines it can produce once effectively trained)
Then no smaller time actor, voice actor, artist, musician, etc, is likely to have any chance defending themselves against the theft of their work for AI purposes.
And, with that precedent set, the legal and financial landscape of art and creativity will have changed in a way that discourages anybody from creating original works for financial gain, because we've systematised the creation of original and derivative works with no legal or financial ramifications.
Cons: the rise of scams, carbon emissions, and, some think, existential risk.
Pros: ??? You can be more efficient at the email job you hate. And it gives you advice on gardening and home improvement so you don’t have to call your uncle.
Cons outweigh the pros
Kinda delegitimizes your cons list too. I'm not going to bother word-by-word checking your work.
People aren't as selfish/petty as you're making them out.
Well. This reaction invites further psychological interpretations, lol.
Never attribute to jealousy that which can be explained by a difference of perspective.
Jealousy can explain any criticism of anybody. That makes it an epistemic hazard; you can always fall back to it, and you can develop a habit of doing so. And then you have deafened yourself. Any substantive criticisms will be lost on you. As will any opportunity to learn from those you disagree with. Your ideas will be hot house flowers, comfortable and safe in their controlled environment, but unable to contend in the wild. Aspire to weeds.
If you choose to do nothing with that information, that's up to you.
If your business model only works if you steal stuff from other people you don’t have a business.
Where can I insert my money to buy every penguin random house book? How do I pay every deviant art artist whatever amount they're owed?
All of deviant art? Probably not. But do you need to train on all of that? You could certainly run ads telling people you’d be willing to pay a small amount to train on their art and let them choose.
Would it be legal to train against the national archives?
Options exist. No you won’t get as much stuff as you do by taking whatever you want, but people are being compensated for their work or at least being given the choice to opt in.
The reason why open-source AI exists at all is that we've always been allowing use of public data - it was okay when Google did it, it was okay when the Internet Archive did it, it was even okay when text translation services used that same data to train their models - or really, that applies to basically anything ML-driven before generative AI.
There's, like, a sea of reasons to criticize OpenAI for - but arguing for extending IP laws even further and calling out opponents for "literal theft" is one of the weaker options that caught on with many people.
ipso facto why does it not exist while spotify buys podcasts and kills them
It's really weird that someone would think that IP law is more important than access to information. Must be some dystopian bubble.
You can frame it however you like to make it justifiable, but I'm not going to get into the mud with you.
They unleashed a massive new wave of spam and scams and garbage onto the Internet?
Because they’re stealing every single bit of content on the Internet and everywhere else they can get their hands on without paying anything for it and then expect to sell it back to us in chewed up garbage form?
They help single-handedly cause MS to blast past their carbon commitments by 30%? And got everyone else into a big race for how many resources they could waste to power AI nonsense that doesn’t even actually work that well for what people want to use it for?
Perhaps the fact that they don’t seem to care one bit about any of the consequences, legal/moral/ethical/economical/etc. caused by what they’ve done as long as they make money?
I don’t have a grudge against open AI. I have a grudge against the AI industry and the way these kind of SV golden boys seem to think they’re immune from criticism.
Why do you think stealing a professional actor’s voice should be OK and immune from criticism? This is horrible.
It’s important to nip that shit in the bud lest it spread.
Stole someone's voice? That's not stealing, let's not pump more power for copyright propaganda even if this case is correct.