I admit that I am biased enough to immediately expect the AI agreement to be exactly what we need right now if this is how Meta reacts to it. Which I know is stupid because I genuinely have no idea what is in it.
Edit: from the linked in post, Meta is concerned about the growth of European companies:
"We share concerns raised by these businesses that this over-reach will throttle the development and deployment of frontier AI models in Europe, and stunt European companies looking to build businesses on top of them."
These regulations may end up creating a trap for European companies.
Essentially, the goal is to establish a series of thresholds that result in significantly more complex and onerous compliance requirements, for example when a model is trained past a certain scale.
Burgeoning EU companies would be reluctant to cross any one of those thresholds and have to deal with sharply increased regulatory risks.
On the other hand, large corporations in the US or China are currently benefiting from a Darwinian ecosystem at home that allows them to evolve their frontier models at breakneck speed.
Those non-EU companies will then be able to enter the EU market with far more polished AI-based products and far deeper pockets to face any regulations.
It's basically micromanaging an industry that European countries have not been able to cultivate themselves. It's legislation for legislation's sake. If you had a naive hope that Mario Draghi's gloomy report on the EU's competitiveness would pave the way for a political breakthrough in the EU - one is tempted to say something along the lines of communist China's market reforms in the 70s - then you have to conclude that the EU is continuing in exactly the same direction. I have actually lost faith in the EU.
How about not assuming by default? How about reading something about this? How about forming your own opinion, and not the opinion of the trillion- dollar supranational corporations?
Let's see, how many times did I get robo-called in the last decade? Zero :)
Sometimes the regulations are heavy-handed and ill-conceived. Most of the time, they are influenced by one lobby or another. For example, car emissions limits scale with _weight_ of all things, which completely defeats the point and actually makes today's car market worse for the environment than it used to be, _because of_ emissions regulations. However, it is undeniable that that the average European is better off in terms of privacy.
The US, China and others are sprinting and thus spiraling towards the majority of society's destitution unless we force these billionaires hands; figure out how we will eat and sustain our economies where one person is now doing a white or blue (Amazon warehouse robots) collar job that ten use to do.
People complain more about cookie banners than they do the actual invasive tracking by those cookies.
Those banners suck and I wouldn't mind if the EU rolled back that law and tried another approach. At the same time, it's fairly easy to add an extension to your browser that hides them.
Legislation won't always work. It's complex and human behavior is somewhat unpredictable. We've let tech run rampant up to this point - it's going to take some time to figure out how to best control them. Throwing up our hands because it's hard to protect consumers from power multi-national corporations is a pretty silly position imo.
Just like GDPR, it will tremendously benefit big corporations (even if Meta is resistant) and those who are happy NOT to follow regulations (which is a lot of Chinese startups).
Not just Meta, 40 EU companies urged EU to postpone roll out of the ai act by two years due to it's unclear nature. This code of practice is voluntary and goes beyond what is in the act itself. EU published it in a way to say that there would be less scrutiny if you voluntarily sign up for this code of practice. Meta would anyway face scrutiny on all ends, so does not seem to a plausible case to sign something voluntary.
One of the key aspects of the act is how a model provider is responsible if the downstream partners misuse it in any way. For open source, it's a very hard requirement[1].
> GPAI model providers need to establish reasonable copyright measures to mitigate the risk that a downstream system or application into which a model is integrated generates copyright-infringing outputs, including through avoiding overfitting of their GPAI model. Where a GPAI model is provided to another entity, providers are encouraged to make the conclusion or validity of the contractual provision of the model dependent upon a promise of that entity to take appropriate measures to avoid the repeated generation of output that is identical or recognisably similar to protected works.
It doesn't seem unreasonable. If you train a model that can reliably reproduce thousands/millions of copyrighted works, you shouldn't be distributibg it. If it were just regular software that had that capability, would it be allowed? Just because it's a fancy Ai model it is ok?
The quoted text makes sense when you understand that the EU provides a carveout for training on copyright protected works without a license. It's quite an elegant balance they've suggested despite the challenges it fails to avoid.
> One of the key aspects of the act is how a model provider is responsible if the downstream partners misuse it in any way
AFAICT the actual text of the act[0] does not mention anything like that. The closest to what you describe is part of the chapter on copyright of the Code of Practice[1], however the code does not add any new requirements to the act (it is not even part of the act itself). What it does is to present a way (which does not mean it is the only one) to comply with the act's requirements (as a relevant example, the act requires to respect machine-readable opt-out mechanisms when training but doesn't specify which ones, but the code of practice explicitly mentions respecting robots.txt during web scraping).
The part about copyright outputs in the code is actually (measure 1.4):
> (1) In order to mitigate the risk that a downstream AI system, into which a general-purpose AI model is integrated, generates output that may infringe rights in works or other subject matter protected by Union law on copyright or related rights, Signatories commit:
> a) to implement appropriate and proportionate technical safeguards to prevent their models from generating outputs that reproduce training content protected by Union law on copyright and related rights in an infringing manner, and
> b) to prohibit copyright-infringing uses of a model in their acceptable use policy, terms and conditions, or other equivalent documents, or in case of general-purpose AI models released under free and open source licenses to alert users to the prohibition of copyright infringing uses of the model in the documentation accompanying the model without prejudice to the free and open source nature of the license.
> (2) This Measure applies irrespective of whether a Signatory vertically integrates the model into its own AI system(s) or whether the model is provided to another entity based on contractual relations.
Keep in mind that "Signatories" here is whoever signed the Code of Practice: obviously if i make my own AI model and do not sign that code of practice myself (but i still follow the act requirements), someone picking up my AI model and signing the Code of Practice themselves doesn't obligate me to follow it too. That'd be like someone releasing a plugin for Photoshop under the GPL and then demanding Adobe release Photoshop's source code.
As for open source models, the "(1b)" above is quite clear (for open source models that want to use this code of practice - which they do not have to!) that all they have to do is to mention in their documentation that their users should not generate copyright infringing content with them.
In fact the act has a lot of exceptions for open-source models. AFAIK Meta's beef with the act is that the EU AI office (or whatever it is called, i do not remember) does not recognize Meta's AI as open source, so they do not get to benefit from those exceptions, though i'm not sure about the details here.
It's a trojan horse, they try to do the same thing that is happening in the banking sector.
By this they want AI model provider to have a strong grip on their users, so controling their usage to not risk issues with the regulator.
Then, the European technocrats will be able control the whole field by being able to control the top providers, that then will overreach by controlling their users.
Kaplan's LinkedIn post says absolutely nothing about what is objectionable about the policy. I'm inclined to think "growth-stunting" could mean anything as tame as mandating user opt-in for new features as opposed to the "opt-out" that's popular among US companies.
Europe is the world‘s second largest economy and has the world‘s highest standard of living. I’m far from a fan of regulation but they’re doing a lot of things right by most measures. Irrelevancy is unlikely in their near future.
> It aims to improve transparency and safety surrounding the technology
Really it does, especially with some technology run by so few which is changing things so fast..
> Meta says it won’t sign Europe AI agreement, calling it an overreach that will stunt growth
God forbid critical things and impactful tech like this be created with a measured head, instead of this nonsense mantra of "Move fast and break things"
Id really prefer NOT to break at least what semblance of society social media hasn't already broken.
The more I read of the existing rule sets within the eurozone the less surprised I am that they make additional shit tier acts like this.
What do surprise me is anything at all working with the existing rulesets, Effectively no one have technical competence and the main purpose of legislation seems to add mostly meaningless but parentally formulated complexities in order to justify hiring more bureaucrats.
>How to live in Europe
>1. Have a job that does not need state approval or licensing.
>2. Ignore all laws, they are too verbose and too technically complex to enforce properly anyway.
I hope this isn't coming down to an argument of "AI can't advance if there are rules". Things like copyright, protection of the sources of information, etc.
Meta on the warpath, Europe falls further behind. Unless you're ready for a fight, don't get in the way of a barbarian when he's got his battle paint on.
As a citizen I’m perfectly happy with the AI Act. As a “person in tech”, the kind of growth being “stunt” here shouldn’t be happening in the first place. It’s not overreach to put some guardrails and protect humans from the overreaching ideas of the techbro elite.
Nit: (possibly cnbc's fault)
there should be a hyphen to
clarify meta opposes overreach, not growth. "growth-stunting overreach" vs "growth (stunting overreach)"
50 comments
[ 2.8 ms ] story [ 70.0 ms ] threadhttps://openai.com/global-affairs/eu-code-of-practice/
Edit: from the linked in post, Meta is concerned about the growth of European companies:
"We share concerns raised by these businesses that this over-reach will throttle the development and deployment of frontier AI models in Europe, and stunt European companies looking to build businesses on top of them."
https://artificialintelligenceact.eu/introduction-to-code-of...
It’s certainly onerous. I don’t see how it helps anyone except for big copyright holders, lawyers and bureaucrats.
Essentially, the goal is to establish a series of thresholds that result in significantly more complex and onerous compliance requirements, for example when a model is trained past a certain scale.
Burgeoning EU companies would be reluctant to cross any one of those thresholds and have to deal with sharply increased regulatory risks.
On the other hand, large corporations in the US or China are currently benefiting from a Darwinian ecosystem at home that allows them to evolve their frontier models at breakneck speed.
Those non-EU companies will then be able to enter the EU market with far more polished AI-based products and far deeper pockets to face any regulations.
What exactly is onerous about it?
Am I the only one who assumes by default that European regulation will be heavy-handed and ill conceived?
And that's the problem: assuming by default.
How about not assuming by default? How about reading something about this? How about forming your own opinion, and not the opinion of the trillion- dollar supranational corporations?
Sometimes the regulations are heavy-handed and ill-conceived. Most of the time, they are influenced by one lobby or another. For example, car emissions limits scale with _weight_ of all things, which completely defeats the point and actually makes today's car market worse for the environment than it used to be, _because of_ emissions regulations. However, it is undeniable that that the average European is better off in terms of privacy.
Those banners suck and I wouldn't mind if the EU rolled back that law and tried another approach. At the same time, it's fairly easy to add an extension to your browser that hides them.
Legislation won't always work. It's complex and human behavior is somewhat unpredictable. We've let tech run rampant up to this point - it's going to take some time to figure out how to best control them. Throwing up our hands because it's hard to protect consumers from power multi-national corporations is a pretty silly position imo.
And consumers will bear the brunt.
One of the key aspects of the act is how a model provider is responsible if the downstream partners misuse it in any way. For open source, it's a very hard requirement[1].
> GPAI model providers need to establish reasonable copyright measures to mitigate the risk that a downstream system or application into which a model is integrated generates copyright-infringing outputs, including through avoiding overfitting of their GPAI model. Where a GPAI model is provided to another entity, providers are encouraged to make the conclusion or validity of the contractual provision of the model dependent upon a promise of that entity to take appropriate measures to avoid the repeated generation of output that is identical or recognisably similar to protected works.
[1] https://www.lw.com/en/insights/2024/11/european-commission-r...
AFAICT the actual text of the act[0] does not mention anything like that. The closest to what you describe is part of the chapter on copyright of the Code of Practice[1], however the code does not add any new requirements to the act (it is not even part of the act itself). What it does is to present a way (which does not mean it is the only one) to comply with the act's requirements (as a relevant example, the act requires to respect machine-readable opt-out mechanisms when training but doesn't specify which ones, but the code of practice explicitly mentions respecting robots.txt during web scraping).
The part about copyright outputs in the code is actually (measure 1.4):
> (1) In order to mitigate the risk that a downstream AI system, into which a general-purpose AI model is integrated, generates output that may infringe rights in works or other subject matter protected by Union law on copyright or related rights, Signatories commit:
> a) to implement appropriate and proportionate technical safeguards to prevent their models from generating outputs that reproduce training content protected by Union law on copyright and related rights in an infringing manner, and
> b) to prohibit copyright-infringing uses of a model in their acceptable use policy, terms and conditions, or other equivalent documents, or in case of general-purpose AI models released under free and open source licenses to alert users to the prohibition of copyright infringing uses of the model in the documentation accompanying the model without prejudice to the free and open source nature of the license.
> (2) This Measure applies irrespective of whether a Signatory vertically integrates the model into its own AI system(s) or whether the model is provided to another entity based on contractual relations.
Keep in mind that "Signatories" here is whoever signed the Code of Practice: obviously if i make my own AI model and do not sign that code of practice myself (but i still follow the act requirements), someone picking up my AI model and signing the Code of Practice themselves doesn't obligate me to follow it too. That'd be like someone releasing a plugin for Photoshop under the GPL and then demanding Adobe release Photoshop's source code.
As for open source models, the "(1b)" above is quite clear (for open source models that want to use this code of practice - which they do not have to!) that all they have to do is to mention in their documentation that their users should not generate copyright infringing content with them.
In fact the act has a lot of exceptions for open-source models. AFAIK Meta's beef with the act is that the EU AI office (or whatever it is called, i do not remember) does not recognize Meta's AI as open source, so they do not get to benefit from those exceptions, though i'm not sure about the details here.
[0] https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:...
[1] https://ec.europa.eu/newsroom/dae/redirection/document/11811...
By this they want AI model provider to have a strong grip on their users, so controling their usage to not risk issues with the regulator. Then, the European technocrats will be able control the whole field by being able to control the top providers, that then will overreach by controlling their users.
Really it does, especially with some technology run by so few which is changing things so fast..
> Meta says it won’t sign Europe AI agreement, calling it an overreach that will stunt growth
God forbid critical things and impactful tech like this be created with a measured head, instead of this nonsense mantra of "Move fast and break things"
Id really prefer NOT to break at least what semblance of society social media hasn't already broken.
What do surprise me is anything at all working with the existing rulesets, Effectively no one have technical competence and the main purpose of legislation seems to add mostly meaningless but parentally formulated complexities in order to justify hiring more bureaucrats.
>How to live in Europe >1. Have a job that does not need state approval or licensing. >2. Ignore all laws, they are too verbose and too technically complex to enforce properly anyway.
Usually minorities are able to impose "wins" on a majority when the price of compliance is lower than the price of defiance.
This is not the case with AI. The stakes are enormous. AI is full steam ahead and no one is getting in the way short of nuclear war.