Among the many recent controversies around AI, there is the claim that to solve a longstanding mathematical problem, OpenAI stole the work of the New York University mathematician Tristan Buckmaster and a colleague, Levent Alpöge, who also is a researcher for Anthropic. I have no ability to assess the validity of the claim, but as I understand it, the argument is that OpenAI had access to work done by Buckmaster and Alpöge. It then built on this work to quickly rush ahead and find a solution, which also had a $1 million prize attached to it.
Insofar as that accurately describes what OpenAI did, this seems like it would be equivalent to having a professor doing a seminar for colleagues and then having one of the attendees do a few additions and claim the work as their own. This would not be viewed as very collegial behavior in a normal academic department.
I’m not sure what claim a professor would have against a colleague under such circumstances. I suppose if their work was properly cited, there would not be a claim of plagiarism. (OpenAI’s work did not directly cite the work of Buckmaster and Alpöge, from my understanding.) I suppose there could be a basis for disputing rights to the million-dollar prize. But the problem seems much greater.
If AI can effectively steal the work of others, and claim it as its own, then it undermines the practice of openness in science that has allowed for enormous progress for centuries. This practice depended on norms, whereby scientists would properly credit the work of their peers.
In the seminar scenario I described, if one of the attendees listened to the presentation and realized the one or two additional steps needed to get over the goal line, they would be expected to raise the issue with the presenter. Assuming the additions were correct, the presenter would finish the paper, acknowledging the contribution or possibly adding the attendee as a co-author, if their contribution was sufficiently important.
But that relies on norms. If the owner of the AI doesn’t share those norms and can freely scrape scientific work wherever it finds it, and then take credit for the product, it will threaten the longstanding practice of openness in science. With the risk of having years of work stolen without getting any credit, researchers will likely keep their work closely held until they have a finished paper.1 They would only share it with a small group of trusted colleagues. Posting a working paper, or having an open seminar, would put control over their work at risk.
However, it is important to recognize the villain in this story. It is Sam Altman and his company, not the AI. Altman and the company are the ones responsible for what their AI does. If it did something they disapprove of, it is their responsibility to correct it. However adept AI may become, it is not the active player; it is the individual or company that set it in motion.
This issue applies more generally. There has been a social media mini firestorm over Texas Attorney General and Republican Senate candidate Ken Paxton using AI to have his opponent, James Talarico, making a series of outlandish statements that he never said. Much of the outrage seems to be over Paxton using AI, rather than the fact that he is lying about what Talarico said.
The issue here is that Paxton is defaming Talarico; the AI is secondary. The AI presumably makes the defamation more effective, but if Paxton were an exceptionally gifted mimic, he could likely also trick some people into believing that he was speaking in Talarico’s voice when making outlandish assertions. If that were the case, it would be every bit as bad as using AI to accomplish the job.
There were also the claims that the girls’ school in Iran, which was destroyed by the US military on the first day of the war, killing 120 children, was selected as a target by AI. That may be the case, but the important point was that the military did not have a system in place to properly review targets. It doesn’t matter whether or not an AI system identified the target. The problem was that there was no system that ensured the places designated were proper military targets.
It is important to keep our eyes on the ball. The bad actors are those who do bad things with AI, not the AI. This is like the husband who gets drunk and beats his wife and then says the whiskey did it. No one should ever accept that story. In the same vein, if a theft of mathematical work did occur, it was Sam Altman who did it and it was Ken Paxton who libeled his political opponent. The AI is beside the point.
Why are we acting like it’s news that generative AI companies are building their core products of the theft and exploitation of other people’s intellectual labor?
Get with the program, it’s their entire business model.
The captured SCOTUS said its okay and no further discussion needed as they needed to hit the tarmac asap because they already booked a spin on their new private jet/gift but absolutely not a bribe.
Of course he did, that doesn’t mean AI is a worthwhile product. It is just a means for rich to steal your life.
If AI can effectively steal the work of others, and claim it as its own, then it undermines the practice of openness in science that has allowed for enormous progress for centuries.
Open source developers have been saying this for years, ever since the first ChatGPT was released. It was laundering open source code, claiming the work of others as it’s own, ignoring the license.
Thank you! It’s so weird to see some pro AI people act like this somehow crossed a line. It’s my understanding that part of the issue is that the AI used its chat logs to get some of this information, but I’ve seen no one explain how that is any different from what was done to people who believed their work was private or protected under a license but ended up having it in a training set anyway. I don’t believe there’s an expectation of privacy when using the plagiarism bot, and my understanding is it does reserve the right to train off your chats.
The concern regarding attribution is also strange to me because it’s possible that the attribution for any specific thing is completely lost during the training process. When people have it generate images or stories or even just a response, it’s not going line by line explaining all the places it stole from. When people talked about AI bootstrapping itself there was no discussion of how we were going to keep track of all the people whose hard work it was building off of.
I’m against AI in general for a bunch of reasons, but the fact they’re upset they got scooped is so minimal in my opinion. Wasn’t AI supposed to solve all these problems for us, faster than we could alone? It’s doing exactly what they said it would, and since they never bothered to care about attribution before I’m having a hard time sympathizing with them now.
I stopped contributing to open source code because of it. I consider it stupid now to contribute to public work because it will be almost instantly stolen and laundered through an AI that breaches the license its published under.
I understand that, but what tools would you use if all open source devs stopped their contributions? there wouldn’t be anything left. we could go back to windows 11, googlified android with all the scam ads of the world, etc. I feel this mindset hurts the open source community more than the ubercapitalist pigs
“guns don’t kill people. people kill people.”
With guns, POW!
I think OP is trying to say that we cannot let “AI” launder the responsibility that real humans have for their actions.
The point (as I understand it) of “guns don’t kill people” is that it is patently absurd to imply that “tools” are wholly neutral implements. The quote is trying to hold the “tool” accountable. While it’s making a seemingly contradictory point to OP, I think it’s more that both can be true at once–OP is trying to highlight the specific problem with “the AI is the one doing all the naughty things”, because Sam Altman et al. would certainly like everyone to think so. I totally agree that “AI” is not merely a “tool” (in the sense of the neutrality so-implied), but that is a separate point.
(re-reading what I just wrote… it kinda sounds like clanker speak?? I swear I’m a real human)
Got it, guns kill AI
In this scenario, everyone is so focused on the gun that they forgot murder is already illegal. We do need gun and AI regulation, but we can’t even put the AI aspect aside for one second to investigate theft and defamation which we do have laws against regardless of the tools used.
If murder is illegal then the people making the tools specifically designed to kill should go to jail.
I agree, but I also think (and argue it is a logical conclusion) that war should then also be illegal (as it uses tools designed to kill) and those who try to instigate it should go to jail. Unfortunately that is a long step from toady’s world.
I genuinely don’t understand how war is legal.
How would you enforce a law against war, and who would it apply to?
Cause if you have enough guys to do a war then you have enough guys to stop you from being arrested.
… with guns.
Old people burning, old people burning, put your hands up!
Thank you Freddy.
… designed specifically for killing
just wait until you find out about watson and crick!
Remind me: it was a female grad student who did the work, right?
It was a male grad student (Raymond Gosling) who did the work for Rosalind Franklin; Franklin reported to Maurice Wilkins; Franklin, Wilkins, Watson, and Crick collaborated for a two-paper submission to Nature which publicized the X-ray evidence for, and the theoretical implications of, DNA’s double-helical structure. The surviving members of this collaboration won the Nobel Prize (it is not generally awarded posthumously.)
This has morphed into a claim that “Watson and Crick stole Franklin’s work” because they’re white men and she’s a woman.
The metaphor of the open seminar is misleading. In an open seminar, the presenter stake a claim to the work, even if incomplete. In this case, the work was in private repositories. A more precise methodology would be the AI overhearing the authors discussing in their office and using that as base for a proof. This is even sneakier because there has been no public claim of the method or results by the authors.
Edited after a commenter pointed out that no sleep + autocorrect = random words
In an open seminar, the presenter pays cousin in the with, even if incomplete.
I can’t figure out what this was meant to say.
Oh wow… it mostly means that I have a small kid that kept me awake most of the night! Edited
This seems like a series of straw-man arguments that conflate separate issues that get separate criticisms.
Yes of course there are a large number of people that criticise any use of generative AI, whether it’s for good or bad. But this doesn’t mean that every criticism involving AI is a criticism of the use of AI, rather than a criticism of the way that it is being used by the person using it. You would have to be extremely inept to make that mistake of comprehension.
In the case of a politician creating fake speeches of their opponent, I can’t believe for a second that the criticism in this “social media mini firestorm” is focusing more on the use of AI than the fact that a politician is committing gross misconduct. If I saw people making that mistake on social media I would assume that it was a deliberate attempt to divert attention from the politician onto the AI tool by the politician’s PR team because again, you would have to be extremely inept not to see the actual source of the controversy.
In the case of the use of AI to find military targets in Iran resulting in the bombing of a school, the criticism is clearly that the person who has ordered this system to be used and everyone in the chain of command who didn’t point out what a bad idea it is, is grossly incompetent (or callously careless, or suffering AI psychosis or all three). No matter how pro-AI you are, everyone at this point should be acutely aware that if you can’t trust an AI to set a calendar reminder and not delete your hard disk, you sure as shit can’t trust it to choose which point on a map of another country you should fire missiles at. To think that this furore was because they used AI at all rather than what they were using it for you would have to be extremely inept.
So what is the motivation of this article? It feels like a completely unnecessary defence of AI as a technology against criticism it hasn’t received as a way to distract from the less defensible criticism that it has received.