> The tech company [Carpe Data] [0] looks at anything publicly available on social media
Keyword being public. I've never understood why people get so upset about this sort of thing. You are responsible for anything you choose to post publicly.
My feelings would be different if it involved fake accounts friending users and seeing items meant only for their "friends."
[0] This article is basically an ad for Carpe Data. May as well link to their site: https://carpe.io/
Just today I was going to go skiing but my foot hurt so I'm home. Let's pretend it was something bigger and I've been home and being paid on an insurance claim. My family makes a new facebook post of last years trip with all of us in the photo and no description or anything. Carpe does the happy dance and sends it to Allstate. Allstate says we are not paying your claim and canceling your policy for fraud.
Now I have to prove the claim, can't get insurance cause I'm on a blacklist and hell knows what else due to an inaccurate photo scraped by a company without official access.
Or you just say "yes, this trip was in the following dates and not during my claim, which you can see from (provided documentation or other posts you made)" (and you can also untag yourself from other people's photos)
Do you tho? Do you "just say"? Out loud? To the room? Does your couch hear you and fill out all the bullshit paperwork, wait hours on hold, go to court?
The problem is you'll never know what were the actual criteria used to deny your claim, it will remain some vague denial reason. Or you will get to know but after much back-and-forth during which it incumbs to you to pony up with that cash flow.
This opacity is just another form of information asymmetry, benefiting the large asset holders and negatively impacting the individuals.
Pretty much everything digital has metadata tagged on it including the time at which the event occurred. Its very easy to prove the timing of the photo in this hypothetical case.
Insurance is a good that adds to the overall wellbeing of society IF it can be properly priced - meaning both fraud as well as incorrect denials of claims damage the insurance companies to properly price coverage and provide protection for everyone.
Why would we dislike more complete information? Economic systems always work better with fully informed players.
The flip side of that is that many services mangle or strip meta-data. Exif data is also easily editable and not all cameras, phones, or software properly populate it. I can see a situation where someone edits a photo before uploading which causes the creation date to default to the date it was edited and not the date the photo was taken.
I also doubt Facebook throws away photo metadata when an image is uploaded.
The image that gets returned in your browser may have none but the original, with all its camera-identifying, time-stamped GPS located gold mine will be intact.
Yeah, I imagine the putatively photographer sites like Flickr preserve it. My assumption is that Facebook saves it, removes it, then allows the image to display. ISTR years ago some controversy about EXIF stalking on FB, but it may have been another service.
Ideally, it is true that COMPLETE information makes an economic system work better..... but the improvement isn't some sort of linear progression, where every bit of more information adds a bit more economic efficiency. Sometimes, getting more information without getting ALL the information actually hurts economic efficiency.
A very easy example of this is wikileaks and the elections in 2016. It appears that wikileaks received information about both the democrat and republican parties; however, they chose to only share the information they recieved about the democrats.
By your assertion, we should not be concerned about this; after all, the information that was leaked was true, and therefore added to our overall body of known information, we should have been in a better place to make an informed decision, right? More information is better!
However, this isn't the case; because the information was selectively released, we only got the information that benefited one side, meaning that our overall understanding of the state of things actually DECREASED as more information was released.
This is the same thing happening here; you know Allstate isn't going to be looking for social media information that HELPS your claim. They aren't going to say, "Well, we were going to deny your claim, but we found from your social media feed that you were clearly really hurt. We have decided to pay out instead!"
Basically, if the extra information you receive is being selectively given by someone who has an agenda, you can be mislead even if all the information is true.
you know Allstate isn't going to be looking for social media information that HELPS your claim.
But from the article:
As for use cases for social media during the insurance claims process, Carpe Data cited two examples. The first one was where a customer was paralyzed. A social media search helped confirm the information more quickly through pictures of the customer in a wheelchair, which ultimately led to the claim being settled and paid out faster.
It's presumably in Allstate's interest to identify fraud, but also to process claims as efficiently as possible, no?
> The first one was where a customer was paralyzed. A social media search helped confirm the information more quickly through pictures of the customer in a wheelchair
This seems highly dubious. Are they really using information like this to decide to pay a claim? It's pretty easy to put someone into a wheelchair and take a video of it.
The flip side of that is that it's pretty challenging to take video of a paralyzed person playing basketball. It seems clear that the use of this is going to be biased in one direction.
Perfect information is impossible. The problem people have with this is that social media information may not enough have enough context to provide anything more than noise or a misleading picture.
> Economic systems always work better with fully informed players.
Except in the current commercial environment, most information is shrouded by the industry and the regulatory processes systematically corrupted. Allstate in particular weaponizes [1] its claims process to low-ball claims, slow legal proceedings and starve-out claimants until they settle. We should be dubious social media information will be used fairly.
Allstate is the only player better informed. The % of Allstate customers that know their social media is being scraped to detect fraud has got to be incredibly small.
If anything, this is asymmetric, which rarely turns out well for the player with less information.
is it really any different than a private investigator (or even a company one) snooping around your residential area and talking to you and your family ... perhaps even seeing the photo in question on your parent's mantle?
I agree that people should be more responsible about what things they choose to post publicly. But networks such as facebook make it really difficult for the average joe (the guy that rock climbs after claiming a back injury) to decipher what is private vs. public. When it's tied to your real-life identity, and potential consequences therein, it should be so simple a toddler could understand.
I don't think this is as much of a problem with the social contract of privacy as it is about the social contract regarding distributed risk. I'm happy to pitch in publicly via taxes and privately via insurance to make sure people are taken care of in the event of a serious injury, because it will probably happen to me eventually and I'd like to be taken care of too. But if you've got a serious injury claim, you should NOT BE ROCK CLIMBING. The idea that the message there is that you shouldn't post public pics of you rock climbing is the wrong message imo.
Reddit, much like this site; offers the potential for anonymity but it is all too easy to inadvertently leak correlation handles that allow for deanonymization.
General rule is going to be the longer you use an identity the more likely you are to taint it with information having direct or indirect deanonymization potential.
But the _spirit_ of social media is to publicly share information for specific types of social interactions. I have always felt that examples like job recruiters researching applicants' facebook content is morally corrupt, although I don't expect it to ever stop.
For a crude analogy, let's say you're having a yard sale in your front yard. Everything for sale is on public display. How would you feel if a company rep from xyz corp decided they could make a better profile of you by judging all of the contents of your yard sale? It's publicly available data, but is it appropriate?
A friend of mine worked as a private investigator for years after his dream of entering law enforcement didn't pan out. His job was mostly investigating insurance claims. He would actually drive his car and park near the home of the claimant for days and take photographs of any activity that would tend to support or disprove their disability claim. He's moved on to another career now, mostly because the career path in that field was not great.
I'm not seeing how this is more invasive. I mean, people are making certain content public. Having a PI parked up the block in a car with tinted windows seems much more invasive and that is business as usual and legal for the insurance industry.
They are expensive to pay out and often real chronic back pain is not easily differentiated from fake chronic back pain by a physical. The diagnosis is made based on patient reported pain.
The easiest way to confirm the diagnosis is observe if the claimant is behaving in a manner consistent with debilitating back pain.
I’m a lawyer and these tactics (for better or worse) are par for the course. Say the claimant in a civil suit makes a claim that due to their injury they can’t engage in normal daily activities leading to a loss in quality of life, example, I can’t lift up and hold my kids or I can’t take out the garbage. Better believe pictures of those things will pop up at a deposition, if the claimant has in fact done any of the above.
I worked a worker’s compensation claim once, where a contractor claimed that due to a back injury he had basically become immobilized, sure enough he had Facebook posts that included zip lining - and was able to subpoena the waiver he signed declaring he had no injuries - and other posts like “drinking getting fucked up and dancing all night at the club” with pictures. So yeah basically people are doing the job for the insurance companies and lawyers that they used to have to hire and pay PIs to do.
It's not more invasive, but it's much cheaper. Allstate used to have to pay a rent-a-cop to park outside someone's house for every dubious claim. Now they can surveil millions of people almost for free. That's a huge difference.
It's not more invasive for the subject but it is more invasive for a population. The limitations imposed by the laborious nature of an investigation meant it could only be applied in cases where there was already a high degree of suspicion. Tools for scraping and classifying images from social media accounts can be economically deployed against every claim. It's logical that if you file a claim someone will scrape your data and store it for future business use.
Not sure why this comment is dead. Car insurance tracking devices are already a thing, with some scoring the safety of your driving and providing suggestions (recording the G-forces attained and determining that hard control inputs are indicative of dangerous driving).
I've had a few friends volunteer for this already; in my limited experience they are typically used only for a period of time (weeks not years) and are typically marketed as "use our tracking device and get a safety discount."
It strikes me as creepy, but I might think differently about them if I needed a discount.
We use facebook all the time in my bank to check on fraud/claims of stolen credit cards.
Once a customer claimed a stolen card but it was used every day at a specific parking lot. Said customer would post "Checking into work at xxx" every day on facebook where the specific parking lot was xxx's parking lot.
Needless to say I don't have a facebook account or post anything using my full name anywhere public.
The scary part will be when the insurance companies buys a direct feed into private posts from facebook or google sometime in the future.
A good faith effort might be to provide two parallel offers: One based on using the social media meta data coefficient and its commensurate reduced fraud rate, a second tier without social meta data coefficient and obviously without the reduced fraud discount.
52 comments
[ 4.4 ms ] story [ 46.9 ms ] threadKeyword being public. I've never understood why people get so upset about this sort of thing. You are responsible for anything you choose to post publicly.
My feelings would be different if it involved fake accounts friending users and seeing items meant only for their "friends."
[0] This article is basically an ad for Carpe Data. May as well link to their site: https://carpe.io/
Now I have to prove the claim, can't get insurance cause I'm on a blacklist and hell knows what else due to an inaccurate photo scraped by a company without official access.
You don't see a problem with that?
https://news.ycombinator.com/newsguidelines.html
This opacity is just another form of information asymmetry, benefiting the large asset holders and negatively impacting the individuals.
Yes; to you. The insurance company just has to figure out if it's cheaper to see you in court, or to pay for your medical claim.
Insurance is a good that adds to the overall wellbeing of society IF it can be properly priced - meaning both fraud as well as incorrect denials of claims damage the insurance companies to properly price coverage and provide protection for everyone.
Why would we dislike more complete information? Economic systems always work better with fully informed players.
- "Remove All Metadata"
- "Remove Location Information"
- "Remove Person Information"
_All_ of these are removed by default and you explicitly have to enable them.
I also doubt Facebook throws away photo metadata when an image is uploaded.
The image that gets returned in your browser may have none but the original, with all its camera-identifying, time-stamped GPS located gold mine will be intact.
Yeah, I imagine the putatively photographer sites like Flickr preserve it. My assumption is that Facebook saves it, removes it, then allows the image to display. ISTR years ago some controversy about EXIF stalking on FB, but it may have been another service.
A very easy example of this is wikileaks and the elections in 2016. It appears that wikileaks received information about both the democrat and republican parties; however, they chose to only share the information they recieved about the democrats.
By your assertion, we should not be concerned about this; after all, the information that was leaked was true, and therefore added to our overall body of known information, we should have been in a better place to make an informed decision, right? More information is better!
However, this isn't the case; because the information was selectively released, we only got the information that benefited one side, meaning that our overall understanding of the state of things actually DECREASED as more information was released.
This is the same thing happening here; you know Allstate isn't going to be looking for social media information that HELPS your claim. They aren't going to say, "Well, we were going to deny your claim, but we found from your social media feed that you were clearly really hurt. We have decided to pay out instead!"
Basically, if the extra information you receive is being selectively given by someone who has an agenda, you can be mislead even if all the information is true.
you know Allstate isn't going to be looking for social media information that HELPS your claim.
But from the article:
As for use cases for social media during the insurance claims process, Carpe Data cited two examples. The first one was where a customer was paralyzed. A social media search helped confirm the information more quickly through pictures of the customer in a wheelchair, which ultimately led to the claim being settled and paid out faster.
It's presumably in Allstate's interest to identify fraud, but also to process claims as efficiently as possible, no?
This seems highly dubious. Are they really using information like this to decide to pay a claim? It's pretty easy to put someone into a wheelchair and take a video of it.
The flip side of that is that it's pretty challenging to take video of a paralyzed person playing basketball. It seems clear that the use of this is going to be biased in one direction.
That being said, I would still be skeptical of the claim that they will always use this information in a balanced manner.
Perfect information is impossible. The problem people have with this is that social media information may not enough have enough context to provide anything more than noise or a misleading picture.
Except in the current commercial environment, most information is shrouded by the industry and the regulatory processes systematically corrupted. Allstate in particular weaponizes [1] its claims process to low-ball claims, slow legal proceedings and starve-out claimants until they settle. We should be dubious social media information will be used fairly.
[1] https://www.huffingtonpost.com/2011/12/13/insurance-claim-de...
Allstate is the only player better informed. The % of Allstate customers that know their social media is being scraped to detect fraud has got to be incredibly small.
If anything, this is asymmetric, which rarely turns out well for the player with less information.
is it really any different than a private investigator (or even a company one) snooping around your residential area and talking to you and your family ... perhaps even seeing the photo in question on your parent's mantle?
Either get serious about your privacy on public networks tied to your identity (see facebook privacy checker: https://www.facebook.com/help/443357099140264/ ), only use 100% anonymous public networks (see reddit: https://www.reddit.com/ ) or stick to a 100% private social network, like I've done (see select: https://mailchi.mp/231daaf5ccef/select-03 ).
Example from mine: https://imgur.com/3EcvT5E
nitpick: reddit is pesudoanonymous, anonymous would be something like 4chan.
General rule is going to be the longer you use an identity the more likely you are to taint it with information having direct or indirect deanonymization potential.
For a crude analogy, let's say you're having a yard sale in your front yard. Everything for sale is on public display. How would you feel if a company rep from xyz corp decided they could make a better profile of you by judging all of the contents of your yard sale? It's publicly available data, but is it appropriate?
I'm not seeing how this is more invasive. I mean, people are making certain content public. Having a PI parked up the block in a car with tinted windows seems much more invasive and that is business as usual and legal for the insurance industry.
Also, I assume you can't PI every case. So social media may help as a pre-selection step for PI-work.
They are expensive to pay out and often real chronic back pain is not easily differentiated from fake chronic back pain by a physical. The diagnosis is made based on patient reported pain.
The easiest way to confirm the diagnosis is observe if the claimant is behaving in a manner consistent with debilitating back pain.
I worked a worker’s compensation claim once, where a contractor claimed that due to a back injury he had basically become immobilized, sure enough he had Facebook posts that included zip lining - and was able to subpoena the waiver he signed declaring he had no injuries - and other posts like “drinking getting fucked up and dancing all night at the club” with pictures. So yeah basically people are doing the job for the insurance companies and lawyers that they used to have to hire and pay PIs to do.
It's not more invasive, but it's much cheaper. Allstate used to have to pay a rent-a-cop to park outside someone's house for every dubious claim. Now they can surveil millions of people almost for free. That's a huge difference.
"AllState wants to be friends on Facebook (accept/reject)"
It strikes me as creepy, but I might think differently about them if I needed a discount.
Requested permission: view your public profile
Needless to say I don't have a facebook account or post anything using my full name anywhere public.
The scary part will be when the insurance companies buys a direct feed into private posts from facebook or google sometime in the future.