Loading summary
Transcript37 lines
- [00:00]
A
This is infocus, the Hindus current affairs podcast.
- [00:08]
B
Hello and welcome to the Hindus InFocus podcast. I am Nivedita Varadharajan. Last week OpenAI, the maker of ChatGPT, filed for an initial public offering in the United States. With a valuation of over $850 billion, it could be one of the largest tech listings ever. This move marks a significant milestone for a company which began in just 2015 as a not for profit research lab. In 2022, ChatGPT was launched and the company changed its structure and attached a for profit entity to it. This change did not please everyone. Elon Musk, who helped funded OpenAI in its early years before having a fallout to the company's leadership, filed a lawsuit in 2024 questioning the change. He argued that OpenAI moved away from its original mission. He accused founder Sam Altman and President Greg Brockman of manipulating him into donating to a not for profit organization and then attaching a for profit subsidiary and then accepting billions of dollars from Microsoft. In May this year, a jury ruled against Musk on procedural grounds, though he said he will appeal. Is also pursuing a separate antitus case against OpenAI and Microsoft. This dispute is not just a very interesting code drama between Silicon Valley billionaires. As AI becomes more deeply embedded in our day to day lives, what does this case tell us about who controls these big tech organizations and how AI should be governed and how these companies should be governed? To discuss this I'm joined by Rahul Singh, who's an Associate professor at Law at the National School of University. To discuss this, I'm joined by Rahul Singh who's an Associate professor at Law at the National Law School of India. Hello professor, thank you for joining us on our board. Hello professor, thank you for joining us today and welcome to the windfocus podcast. Can you please tell us what the case is all about and what did the jury say on the 18th of May this year? Rahul sir, thank you so much for joining me today.
- [02:13]
A
Thank you Nivedita for having me.
- [02:15]
B
So sir, let's talk about the case itself before we go on to what the case actually means for for all of us. What are the allegations of Elon Musk against to OpenAI and why did the jury reject his claims?
- [02:30]
A
So Elon Musk's claim against OpenAI was twofold. First, because way back in 2015 he had given money to OpenAI and he claimed that that was given under his belief that OpenAI will have a charitable purpose. It appears because jury trials don't have particular opinions and Judgment. So from secondary sources, it appears that in 2015 Elon Musk had given $38 million. In today's conversion rate in India, that will be about round about 400 crore rupees. Now his claim in this jury trial was that given the change in purpose of OpenAI, that instead of being a not for profit, it's changing the purpose to for profit. So Elon Musk's claim is that, you see, this was a breach of trust, or if not a breach of trust, it's certainly an unjust enrichment because money was taken under false pretense. So simply put, these were the two allegations or claims that Elon Musk was posing before the jury.
- [03:49]
B
Why did the jury reject the claims?
- [03:52]
A
So the way jury trials go, I mean they don't have to give reasons, it's very unlike a judgment. But nevertheless, what it appears is that the jury was convinced on a question of fact, which is in order to make a legal claim you must approach a forum within a certain period of time. So what we in India call limitation act, the Americans call it statute of limitations. The jury found that both of these claims came in very delayed. There is some difference in time period for the two claims. For one it is three years and for the second it is two years. But regardless of that time difference, jury found that in 2026, right. A decade has passed. So therefore this is not a claim that can be sustained. So in that sense the, the verdict itself is technical and it can be seen as procedural.
- [04:44]
B
Elon Musk also filed another suit which this jury did not see. I filed. He filed an anti trust suit against OpenAI and also Microsoft. What is that case about?
- [04:58]
A
Right, so that is a slightly different case in the sense that the claims change. And there is some overlap though because the Federal Trade Commission did file a motion interpreting the meaning of unfair competition and that filing is in public domain by the way. So while the jury's verdict is not because in the sense they won't have a judgment, but the FTC's motion that is available on the FTC and DOJ's website. And that claim is slightly different. I mean in terms of let's say what Americans call antitrust and we in India call competition law, the idea there is to figure out if the manner in which OpenAI is functioning, given all of these investments that let's say OpenAI may have in some other companies, including Semiconductor Company, so are they being intertwined in such a manner that all of them control each other and therefore may not be acting independently? They might be acting in a manner which is interdependent. Now, this jury's verdict has no implication on the Federal Trade Commission's claims, although the judge reportedly said something which seems like for the judge, it may have some implication, but it will not have a bearing on whether ftc, the Federal Trade Commission or the Department of Justice independently decides to bring any antitrust suit against OpenAI.
- [06:27]
B
Elon Musk has said that he gave money to OpenAI as a charity because it was a charitable organization. But there are claims that he also wanted OpenAI to have a for profit entity. So how does, how does the jury look at both of them?
- [06:44]
A
It looks like that's exactly what went against Elon Musk, because the way these jury trials go, and it's quite interesting and I would want to draw a comparison with Indian position. So in Indian context, let's say in the famous battle between Cyrus Mistry and Ratan Tata, when the battle was going on at the tribunal in Mumbai, the tribunal had in fact lamented that I don't have this ability to call Mr. Ratan Tata and ask him questions, although the claim itself was inaccurate as per Indian law. But the NCLT Mumbai judicial member laments and laments in writing in the judgment. In that sense, if we compare and contrast with jury trial, it looks like both sides placed their faith before the jury. They didn't have an ego. They turned up both Elon Musk and Sam Altman, they were present in court. In fact, from some of these tellings in US journals and all, it appears that there were journalists who were in fact reporting the manner in which Sam Altman and Elon Musk were behaving with each other in the wardrobe rather than the content of the. But you see, I mean, for us being in India, I think the most exciting bit or interesting bit is that it seems like both parties place their faith in bunch of nine folks who have no training in law because jury trial is trial by a peer group. And Indian context tribunals, which claim to be experts, they lament that we are not able to get these, let's say controllers of large companies in the same room and maybe ask questions, right? So in that sense, you see what is going on is this Elon Musk's claim that this charity and all. So this, these are all based on facts. And Yuri sees the evidence, the evidence here is WhatsApp text or some social media text, and they find that in fact the timelines are slightly, I think blurry. But between 2015 and 2017, it appears that at some point of time, Elon Musk had expressed his own desire to control OpenAI. And this fact, it appears that it goes against him because we can only say it appears because ultimately the jury decides on a technical ground or limitation. And therefore this idea, insofar as it indicates that Musk was aware of, let's say OpenAI's idea of converting to for profit, that happened way before and therefore by that time I suppose Limitation act has passed only to that extent. Possibly it is relevant, but the inference is based on text messages, maybe other kinds of evidence which they came across in the courtroom. Because see both, the moment they land up in accord, they are also subjected to cross examination. And therefore the fact situation which comes across from, let's say during the trial is to indicate that Musk was aware that there is a possibility that OpenAI may take a different turn altogether. It may become for profit.
- [09:46]
B
So that only is very controversial, right? Because OpenAI took a lot of money from a lot of people and it also used a lot of resources claiming that it was a charity, a not for profit entity. And therefore it can use these open sources from the Internet to its AI models. Isn't that a problem now that they are now converting into a for profit company?
- [10:09]
A
Okay, it's. So I think there is also, I mean one will have to see what is the rationale as well. Right, So I think in the. So it again comes across during the trial itself, it appears that OpenAI's claim is that at the stage at which this, let's say large language models are now at or even let's say AGI, the Artificial General intelligence.
- [10:35]
B
Right.
- [10:35]
A
Of course we must use these terms more carefully. But nevertheless, the stage at which the company is. It appears that there is a need for capital and but for capital, the next stage of the company might be jeopardized. So in that sense, while I suppose jury takes a technical sort of a ground on limitation, one can infer that jury probably didn't want to come in the way of the next stage of OpenAI's growth. If OpenAI needs capital, then it must have some ways to access that capital.
- [11:06]
B
Okay, so can you help us identify the difference between a charity and a not for profit company here?
- [11:12]
A
Okay, so I will give Indian examples so that our audience is largely based in India, so that we can appreciate maybe more so under Indian company law as well, it's possible to set up a company for charitable purpose. It doesn't mean that it's not a company, it's a company with charitable purpose. A good example is National Payment Corporation of India, the company which runs UPI infrastructure through BHIM and all. Now that company is set up for charitable purpose. Now there are certain restrictions that automatically comes along with being a charitable purpose company which is for instance, you can't use dividends to for instance return capital to your investors right? Now it doesn't necessarily mean though that you cannot change your company from a charitable purpose to something else. The process in India is slightly different. It requires government approval and all, but a regulatory process in the us so it's not in doubt that the regulatory process in the US has been followed by by OpenAI, whatever that regulatory process is. And therefore in that sense the jury is being cautious that if a regulatory process which permits OpenAI to change, let's say its company's purpose from charitable to for profit, if that has been followed, then there could be maybe some other, for lack of better term, other ulterior purposes that a claimant here has, right? Like for instance, asking this technical question as to why are you so late implies that the claimant was not able to find any credibility with the jury, right? So it implies that, let's say hypothetically if some other claimant comes today, right, not musk, and that other claimants is somehow able to convince that. I was never aware that there could be a for profit turn off event. So hypothetically it may mean that that legal question is still open.
- [13:12]
B
So how is OpenAI structured today? So when we say a nonprofit entity controlling a for profit entity, it sounds a lot like something which, which is a case of model for Tata trusts. There's Tata Trust which controls the Tata group of companies. It's similar to, may appear to be similar.
- [13:30]
A
I mean some of these are idiosyncratic to us. The idea is that you house for profit in one corporate entity so that you will be able to attract investors. Now that's not the purpose for Tatas. Tatas don't necessarily need investor interest. The purpose behind Tata structuring and as it is, this itself is undergoing a lot of controversy right now because Reserve bank of India is also nudging or it appears that RBI is nudging Tatas towards going public, right? So the purpose behind Tata trust structure might be substantially different from the OpenAI as a company structure. But the idea is for us to appreciate that because corporate law deals with different kinds of companies, right? Maybe private company, public companies for profit, not for profit. Therefore there are certain minimum rules of the game that the company's act itself provides. So as of now, until Reserve bank of India categorically mandates for the tatas. It's a permissible way to control your corporate entity through a trust. Similarly, in the US context, if the structure is to merely have a subsidiary and then some holding companies will control it, then that is what the US law is enabling them to do. So the two purposes might be different. But since corporate law has to deal with multiple forms of companies, it has to say something more broadly. And therefore each of these questions which is arising as a dispute is being left to either jury in the US or in Indian context. It will go to NCLT and NCLAT to decide whether something is legal or illegal.
- [15:07]
B
And how ethical is it? Is it even wrong to get an ethical question for OpenAI because it started to take money saying that it's going to be a not for profit entity and now it's converting.
- [15:21]
A
So I mean, the thing is, let's assume hypothetically that it is unethical. The question would be can we enforce it in a forum? You see, in the absence of enforceability, the way we frame a question itself becomes quite contested. If it is unethical, then there should be a forum, there should be a law where somebody is able to enforce it. And if they are able to, then it is just illegal. Now we may find it unethical, but it's merely just illegality, right? Because in the absence of enforcement, you can see, I mean any, any party, and there is a lot at stake here, right? Any party will have perverse incentives to try and find a workaround. The law or the letter of the law. Right. And therefore, I think it's critical to appreciate that a forum is needed, an institution is needed to enforce whether we frame that as ethics or legality.
- [16:17]
B
And how will it change in depending upon how we frame it. If we frame it as legality, there will be strict punishments and checkpoints for all of that. What if it's the other way? What happens if it's a moral question?
- [16:29]
A
Right. Look, I mean, so maybe as a, maybe a simple analogy, right? I mean, look, unless you are in royalty, divorce is permissible in marriages. Now in that sense, let's say this is just some bad relationship, right? Musk is saying I thought X and Sam Altman is saying I thought Y. Right? You see, even marriages permit divorce and it might be unethical to divorce, right? But nevertheless, it's a legal way to deal with breakdown of relationship. Look, while I think much of media is focused on personalities and it seems like persons are quite important here, but I think it's also equally important for Us to keep an eye on what is at stake for rest of us in the society. Given that practically India is nowhere in the AI game because we don't really have a product now. Therefore the future of AI, whether OpenAI will create that future or whether XAI will create or by the way, I mean there are also two other entities, right, which are probably watching this very carefully. One is Anthropic. Just today's newspapers report that Anthropic's valuation has now surpassed opening ads today. And the other one is actually the elephant in the room, Google through Gemini and therefore between these four entities, whoever succeeds is going to write the rules of the game. And therefore the fact that in that sense, let's say in between Musk versus Altman, let's say Musk today might feel that he has lost, but he's appealing by the way. I mean he has stated very clearly he's going to appeal. And it seems like he also finds a way to win in appeals in the US which is very rare, mostly in the US context. Appeal tribunals or appeal courts don't get into questions of fact. But at least in one case Musk has managed to win, which was his compensation case. And therefore we shouldn't underestimate his ability to persuade an appeals court. But for us sitting in India, I think the consequences and the questions should be seen from a slightly different perspective, which is that let's say while these two entities are caught in some kind of a battle, right, Is there an opportunity for changing the structure of the game? Like what I'm suggesting is what does it mean for India? I mean, is there an opportunity for some Indian companies that while they are busy fighting, is it possible for somebody to scale up and come up with a maybe? I mean, it's harder to, let's say start competing today, but at least a minimum viable product which can stand in front of rest of these companies.
- [19:13]
B
So how do we do that? We are nowhere in the picture. We all use ChatGPT, Gemini and all of the others, but we don't have our own Indian version of it. So how will it play out here then?
- [19:26]
A
Now see, I mean that's where we should think about what kind of investment is required and who are the, let's say funders in Indian context who may be in a position to make an investment, right? Like these numbers in that sense give a sense. In 2015 it took, let's say 400 crores. If we extrapolate it with inflation and all, how much will it take today? Maybe if you say 10 times 4,000 crore or even double princess. So that's $1 billion or so, 9,000, 10,000 crore. Now I think the chief economic advisor to the government of India has lamented a couple of weeks ago that he is very frustrated that Indian companies make profit but they make no investments. I'm just citing him. And therefore. And who, which are these companies who are capable of bringing in a capital contribution of let's say $1 billion today? Now those are the companies that the chief economic advisor has in mind. Presumably if they are profitable, they are making profit, then why is it that they are not investing in creating something which may help the Indian society? I think that may be an ethical question because see, we can't really go to court and force large Indian companies to make an investment. But most certainly the chief economic advisor can issue an appeal. And I suppose most certainly we can say that, you know, because it requires high capital investment, not because of any other reason. Hypothetically, a startup can also do certain things, but startups won't have access to a billion dollar. In fact, startups are competing to become a unicorn which is a billion dollar valuation. So if to start with you require that much amount of money, one can only hope that these large companies whom chief economic advisor is saying that they are not investing, they pay some heat to it.
- [21:15]
B
So I like to come back to what this fight means for governance of artificial intelligence before we go on to what it means for India.
- [21:24]
A
Right.
- [21:25]
B
Can we say that this case, it's going on appeal, but will it have, will it be a precedent to how legal battles are taught in the US about, of how AI is controlled?
- [21:37]
A
So most certainly, I think the kind of limelight, the kind of spotlight that this case had when the presidential value in the sense that of. Of course now if somebody else was thinking of filing a similar claim, they would be more careful, right? But at the same time the governance I think is twofold. The governance question like this, OpenAI's commitment to, let's say eliminationary or charitable. Now that's an internal governance, right? But no society must rely on, let's say self promise or self serving promise made by a company, right? And us is not, I mean it's very clear that the Federal Trade Commission still remains very interested in finding out what's happening in the antitrust aspect, right? And therefore the external aspect of governance is certainly not immediately impacted. It may just mean that the regulators or the agencies, as you call in the U.S. agencies in the U.S. might have a bigger task of finding evidence, right? Because the agencies don't decide. Their agencies have to convince the court and therefore the task is slightly higher, more burdensome, they'll have to find evidence to ensure that the allegations against any of these AI companies withstand the scrutiny of the courts. But I think what is going on simultaneously is that there is also some international push, given the kind of impact that AI is supposed to have. I think that's the biggest question of our times anyway, right? What how is AI going to impact? Like I see everybody has suddenly turned a Luddite friendly, including the reporters who write, otherwise known for writing long reads and all, and everybody refused to read them. And they write well. But the undertow, the subtext these days is being a Luddite, right? And because of which I think every other country as well, I mean prominently eu, because traditionally what has happened is that EU has managed to do certain things which rest of the countries haven't, whatever may be the reasons, right? Like Anu Bradford called it the Brussels effect. But how EU may respond to, let's say, governance of the AI, because one part is within US itself, how FTC DOJ might respond. The second part is how you might respond. And you might want to have a broader coalition of nations, right? Perhaps including India. Perhaps one can only hope that India is also on their radar, certainly most certainly China, when nobody is able to ignore China. Right. So it's also a question of how rest of these countries will respond, right? In the governance question, I don't think he who is going to leave it to what US is doing at the FTC level or the DOJ level, I'm sure they will come up with their own set of ideas as to how AI governance must take place.
- [24:33]
B
So is this only about regulating in the form of rules or should countries legislate on it? Can you legislate on entities that operate in another country or are based out of another country right now?
- [24:47]
A
So there is a long standing doctrine called the extraterritorial effect. Now, so long as there is some effect in your country, you can regulate any entity, including India, right? So Indian competition law also has a section which will say moment there is extra territorial impact, let's say in India, then India will regulate. Similarly, EU has a similar doctrine. All countries US has the same doctrine and therefore the effect itself will mean that EU's jurisdiction is not ousted. I think the broader question is whether because see it's very clear today, politically or geopolitics has clarified that every other country except US is really a middle part. And the only way middle parts can act is to act together. None of them individually have sufficient sway. And therefore the larger question is whether these middle parts are willing to negotiate with each other or are they going to just be egotistical and not negotiate with each other and then do certain things which are fragmented.
- [25:48]
B
It's quite safe to say that this dispute will have much larger implications in the world than just Musk versus OpenAI versus Microsoft. We see that Microsoft is now a partner with OpenAI.
- [26:02]
A
And most certainly, I think the one part is it's clear today because the way Anthropic itself has moved ahead within just last one year and the valuation and I don't know if you are aware of the history, I think Karen Lau's book that Empire of AI, it outlines it that the founders of Anthropic, they had a philosophical disagreement. They felt that the commitment to, let's say, social purpose is weak. Therefore they left OpenAI saying that we will have a stronger commitment to social purpose, including the spat which happened between Department of Defense in the US versus Anthropic. Now again, see, that is internal at this stage. Anthropic is taking a particular stance.
- [26:53]
B
Now.
- [26:54]
A
It's not really a guarantee that let's say 10 years later they will say something similar. But even if we see it as a product differentiator or them trying to create a usp, it certainly bodes well for competition in the sector. Right. Look, I mean, a duopoly. So monopoly is the worst. Duopoly is worse than, let's say, oligopoly. Even if there are four entities and it's an oligopolistic market, it's certainly better than just OpenAI calling the shots, or just XAI calling the shots. I think the larger question is whether something is going to come from China or the EU perhaps, I mean, as a moonshot, India. And therefore the market is not just oligopolistic, but can also become more competitive.
- [27:38]
B
That's such a big question. And that's the note we're going to end on today. Thank you so much for joining us today, sir.
- [27:47]
A
Thank you. This episode was edited and produced by Sharmada Venkata Subramanian. If you liked it, do subscribe to Infocus by the Hindu on Spotify, Apple Podcasts or wherever you get your podcasts from. Thank you for listening.