FTC Chair: AI Models Could Violate Antitrust Laws (thehill.com) 42
An anonymous reader quotes a report from The Hill: Federal Trade Commission (FTC) Chair Lina Khan said Wednesday that companies that train their artificial intelligence (A) models on data from news websites, artists' creations or people's personal information could be in violation of antitrust laws. At The Wall Street Journal's "Future of Everything Festival," Khan said the FTC is examining ways in which major companies' data scraping could hinder competition or potentially violate people's privacy rights. "The FTC Act prohibits unfair methods of competition and unfair or deceptive acts or practices," Khan said at the event. "So, you can imagine, if somebody's content or information is being scraped that they have produced, and then is being used in ways to compete with them and to dislodge them from the market and divert businesses, in some cases, that could be an unfair method of competition."
Khan said concern also lies in companies using people's data without their knowledge or consent, which can also raise legal concerns. "We've also seen a lot of concern about deception, about unfairness, if firms are making one set of representations when you're signing up to use them, but then are secretly or quietly using the data you're feeding them -- be it your personal data, be it, if you're a business, your proprietary data, your competitively significant data -- if they're then using that to feed their models, to compete with you, to abuse your privacy, that can also raise legal concerns," she said.
Khan also recognized people's concerns about companies retroactively changing their terms of service to let them use customers' content, including personal photos or family videos, to feed into their AI models. "I think that's where people feel a sense of violation, that that's not really what they signed up for and oftentimes, they feel that they don't have recourse," Khan said. "Some of these services are essential for navigating day to day life," she continued, "and so, if the choice -- 'choice' -- you're being presented with is: sign off on not just being endlessly surveilled, but all of that data being fed into these models, or forego using these services entirely, I think that's a really tough spot to put people in." Khan said she thinks many government agencies have an important role to play as AI continues to develop, saying, "I think in Washington, there's increasingly a recognition that we can't, as a government, just be totally hands off and stand out of the way." You can watch the interview with Khan here.
Khan said concern also lies in companies using people's data without their knowledge or consent, which can also raise legal concerns. "We've also seen a lot of concern about deception, about unfairness, if firms are making one set of representations when you're signing up to use them, but then are secretly or quietly using the data you're feeding them -- be it your personal data, be it, if you're a business, your proprietary data, your competitively significant data -- if they're then using that to feed their models, to compete with you, to abuse your privacy, that can also raise legal concerns," she said.
Khan also recognized people's concerns about companies retroactively changing their terms of service to let them use customers' content, including personal photos or family videos, to feed into their AI models. "I think that's where people feel a sense of violation, that that's not really what they signed up for and oftentimes, they feel that they don't have recourse," Khan said. "Some of these services are essential for navigating day to day life," she continued, "and so, if the choice -- 'choice' -- you're being presented with is: sign off on not just being endlessly surveilled, but all of that data being fed into these models, or forego using these services entirely, I think that's a really tough spot to put people in." Khan said she thinks many government agencies have an important role to play as AI continues to develop, saying, "I think in Washington, there's increasingly a recognition that we can't, as a government, just be totally hands off and stand out of the way." You can watch the interview with Khan here.
Interesting approach for click-through TOC creep. (Score:2)
Happy to see FTC doing something about TOCs changing under antitrust law; I think that is an avenue that is winnable.
yeah but the election (Score:1)
Donald Trump is quite likely to be president in less than a year based on polling. He owns a social media company that is losing tons of money and would probably love to sell data to AI companies. Also I would describe his policies during his last term as being more of a corporate-rights-advocate than a consumer-rights-advocate.
Re: (Score:3, Insightful)
Polling also indicated the "reds" would win half of the special elections since the abortion ruling.
I think they've one 2 or 3 of over 20 special elections.
I think there are nearly 22 million new female voters age 18-30 who are committed to voting blue that are not being polled. And they are in every state.
Re: (Score:2)
Scary: an AI trained on DJT's tweets. Just when you thought the enshitification had hit a maximum...
Re: (Score:2)
I am no fan of Trump at all but I think that would be near impossible to do. The man really is the greater poster of all time, he basically shitposted himself to the Presidency, that's hard to top.
And one of the reasons people like DeSantis and Vivek fell completely flat against him is he just can't be replicated and attempts to mimic his style are immediately obvious. The man is a true a true mutant and I don't mean that in a pejorative way, it's just a fact.
Re: (Score:2)
...he basically shitposted himself to the Presidency, that's hard to top.
Mafia Don learned to bully by abusing the judicial system and being able to pay/con enough lawyers to do his bidding. His shamelessness over social media hasn't been his only trick. His judicial techniques have helped keep him out of prison so far. Like, who else can rack up about a dozen counts of contempt of Federal court while still enjoying personal liberty out and about on US streets following a pre-trial arrest?
I submit to you Mafia Don's latest example of accusing Biden and the FBI of trying to assas [youtube.com]
Re: (Score:2)
Yeah it all ties into his showmanship both in person and online, which I give 80% of his success, even more than his judicial things which I personally think he just stumbled into because he's surrounded by crazies. Trump is an example of a decades long audience capture combined with complete NPD.
Re: (Score:1)
I'm a Trumper, but I want this to be known.
GPT-4o responds (Score:2)
Lina Khan's strongest argument concerns companies retroactively changing their terms of service to incorporate personal content into AI models, which she views as a potential violation of user trust and privacy. However, this concern overlooks the fact that users generally have the opportunity to review and reject these updated terms, maintaining control over their data. Moreover, regulatory frameworks and data protection laws already in place, such as GDPR in Europe and CCPA in California, provide stringent guidelines and enforcement mechanisms against such practices. These laws ensure that any significant changes to data handling practices are not only transparent but also require active consent from the users, thereby safeguarding consumer rights and interests effectively. This framework not only mitigates the risk of unfair competition but also reinforces consumer autonomy, challenging the premise that users are left without recourse.
Khan's concern about companies using proprietary or competitively significant data to unfairly compete appears to assume an environment where such activities are unchecked and rampant. However, this overlooks the robustness of existing antitrust regulations designed to curb such misuse. Most jurisdictions, including the U.S., have established legal standards and regulatory bodies that actively monitor and take action against anti-competitive behaviors, including misuse of data to edge out competitors. Furthermore, these laws compel companies to operate transparently and fairly, ensuring that any use of data is compliant with both competition laws and data privacy regulations. By enforcing these laws, regulators can prevent the monopolistic use of data, thus maintaining a level playing field and protecting smaller entities from being unjustly pushed out of the market. This framework effectively addresses Khan's concerns by providing substantial protections against the issues she highlights.
Re: (Score:2)
overlooks the fact that users generally have the opportunity to review and reject these updated terms, maintaining control over their data.
I see chatgpt has mastered the art of lying.
Re: (Score:2)
Kinda sus GPT just sidestepped the first point about data scraping, in my opinion the TOS thing is actually valid but also the weakest thing here. The data scraping argument I think is where a real crux of the entire issue lies. It's one of those things where it by some measure this practice is not already illegal it probably should be. Multibillion companies should not be able to play the "ask for forgiveness rather than permission" game.
Re: (Score:2)
Data scraping, often depicted as a clandestine corporate maneuver, indeed merits scrutiny when discussing consumer rights and competitive fairness. Yet, the condemnation of this practice across the board ignores the diverse nature of data scraping and its legitimate applications. In cases where data is publicly accessible, scraping serves as a vital tool for innovation, facilitating advancements in various sectors by gathering large datasets essential for developing robust technologies and services. The primary issue isn't the act of scraping itself but how and what data is used for. Regulations like GDPR and CCPA already provide frameworks that require consent and transparency, targeting precisely the misuse of private or sensitive data. Therefore, the solution isn't to vilify data scraping as a whole but to enforce these regulations rigorously, ensuring companies use data responsibly and transparently, thus preventing abuse without stifling innovation. By focusing on compliance and ethical use, we address the core concerns without casting an unwarranted pall over a practice foundational to technological advancement.
Re: (Score:2)
Nice to know GPT supports self enforcement of it's training model it sounds because under it's own criteria it should be allowed to train for free on others data.
Also do t act like I should have known when you didn't disclose that. And also paste your prompt, that's practically just as important.
Re: (Score:2)
The prompt history on the fist post was a little complicated, or I would have posted it.
Re:data scraping (Score:3)
People need to understand that the worldwide web is about sharing (publishing) your information.You don't want it out there, don't put in on there.
Re: (Score:1)
I think the issue is not so much being scraped but being scraped by a fir profit entity with a somewhat explicit purpose to displace your content and from examples enable people to effectively steal it.
"GPT, paint a picture in the style of this artist I like since you scraped all their content and can now do that without ever having asked that artist. "
Legality aside it's a thorny issue.
Re: (Score:2)
Re: (Score:2)
A human no, not at all. A corporation training thousands of humans? Probably more so.
I see it as sort of the difference between me painting a copy of painting or just photocopying it.
What is unfair? (Score:3)
Do these same companies not have access to AI models like everyone else does?
Yes but (Score:2)
Yes but some are better at it than others, and that's unfair.
At some point... (Score:2)
We need to remember that the primary purpose of copyright and IP is to cause more people to disclose their creations to the public and inspire the generation of content.
If at some point this completely artificial legal concept of IP is hindering the mass and ubiquitous creation of content to protect a few horse and buggy drivers the answer is to abolish it rather than block progress.
Re: (Score:2)
You're thinking of patents, where that is certainly the case: to encourage innovation instead of trade secrets.
People have always published books and articles. It has only taken off when the printing press was invented and improved.
Copyright law is just codification of proper conduct towards your fellow man.
Re: (Score:2)
No, I'm thinking of both. Patents were to encourage disclosure and copyright to inspire the generation of content.
Ideas copy naturally and there is no innate restriction to copy. There is no sort of 'proper conduct' towards your fellow man in artificial copy restriction. This notion people are somehow entitled to control copying and distribution of something merely because they managed to scratch out something like it first is very modern and a function of living in a post copyright world. We probably would
Re: (Score:2)
Copyright has always been about control and/or profit, and usually focuses on limiting dissemination. Sometimes for the purpose of censorship, and other times just to make a buck on it. Even in Alexandria, where the law was that you had to let them copy your books, people had to pay to get access to those copies.
I agree though with your final paragraph. I think that's a pretty reasonable system to work under. Who knows what the new paradigm will look like? Will the AI software be refined to make sense relia
Re: (Score:2)
"Copyright has always been about control and/or profit, and usually focuses on limiting dissemination."
I'm referring to the history in the US where freedom of speech and the press were held as core values by the founders. The natural state of nobody having a right to control or even necessarily profit from intellectual property is recognized and sans artificial constructs like copyright everything is in the public domain. Copyright was seen as balance so people could print and distribute creative works whil
Re: (Score:2)
I'm referring to the history in the US
History doesn't begin in the US, and the history of US law doesn't either. It didn't just spring out of nowhere.
where freedom of speech and the press were held as core values by the founders.
Unless the people involved were nonwhite or women, of course.
Re: (Score:2)
"History doesn't begin in the US, and the history of US law doesn't either. It didn't just spring out of nowhere."
That contradicts [or compliments] nothing I said. Is there some relevance to the discussion at hand? We are discussing intellectual property in the United States and copyright law sprang out of Benjamin Franklin; and if we are throwing out asides so did paper entities in the form of trusts.
"where freedom of speech and the press were held as core values by the founders.
Unless the people involved
You can't have it both ways (Score:1)
The current late capitalist system is broken, I think most people would rather see a push for a totally automated post-scarcity society. Sure the transition is painful but short.
https://fortune.com/2023/11/06... [fortune.com]
Re: (Score:2)
No one asked for an AGI.
It's a science fiction trope for a reason, and the idea is about as hopeful as it is feared. Lots of people want AI. For some reason they imagine that it will be rational. But it's not clear that it's rational for it to share our goals, so it's also not clear that's even desirable. Maybe it decides to optimize us out of the picture.
Still, lots of people fantasize about and desire AGI to solve some problem they think we can't solve some other way, or just to leave a legacy that will last longer than humanity
The enshifitication continues. (Score:1)
FTC - Taking publcly available text and repeating it is anti-competitive.
And yet, nobody saw a problem when Google digitized a ton of books.
You can't use someone name, image, or likeness without paying them. That's a viotation of (no law ever).
Truly, what's next, age restrictions on library books? Repeal of "we didn't write it so we shouldn't be sued for it (Sec. 230)?
All these federal oversight organization really do is drive prices up, which means consumers pay more, and their rich
lobbyist-full friends
Re: (Score:3)
If you read a book on plumbing and become a plumber, you owe the author of that book money everytime you fix a sink or unclog a toilet. Oh you went to trade school? You owe your teachers.
Re: (Score:1)
That's a viotation of (no law ever).
It's a fundamental component of torts law, not just generalised, but with specifically recognised legal wording on the books in more than half of US states, but thanks for playing.
This is dumb (Score:2)
The government (yours and everyone else's) and secret entities will have access to all your information and train their AIs on it. Meanwhile if AI is "banned" .. it only means the rest of us won't have access to it. You really want that lop-sided situation? Publicly available information is public as long as someone isn't reproducing it exactly .. it's fair game. My AI is gonna train off it, so fuck off. Learning is different than copying. Do you owe the authors of your textbooks and every stackoverflow art
Re: (Score:2)
Shooting ourselves in the foot (Score:2)
I am not sure why, but the regulators are too keen on crippling the new bugging AI industry before it really becomes mainstream. We just had only a few companies with usable offerings, barely above toy level. Yes, let's be honest, ChatGPT and CoPilot are good, but they are nowhere near self sufficient, nor they have a valid business plan.
Open source on the other hand, is even at a larger disadvantage. Not only they depend on large companies like Meta or Google to release model "weights", any "fine tuning" w
Losing credibility (Score:3)
Is this the same FTC chair that has lead the repeated taking of companies to court and losing on facts of law? How about the FTC actually focuses on dealing with mergers and acquisitions rather than op-eding on the applicability of a law in a very nebulous way.
I'd suggest she go open a law book, but I'm afraid she may ironically be breaching anti-trust laws by learning something from it.
Yet more fishing by these idiots (Score:2)
Universities in violation of anti-trust laws (Score:2)
Students are trained using news websites, artists' creations, others research, and internet data without owner or users consent⦠the universities call this education but it is clearly schools profiting from others work and it must be stopped!
Can Khan (Score:2)
According to Fuhrer Lina, it turns out that anything she doesn't like is a violation of anti-trust laws.