OpenAI reversed its restructuring plans. Critics aren’t cheering.

Summary

POLITICO Tech: OpenAI Reverses Its Restructuring Plans. Critics Aren’t Cheering Release Date: May 8, 2025

In this episode of POLITICO Tech, host Stephen Overle delves into the latest developments surrounding OpenAI, the organization behind ChatGPT. The episode critically examines OpenAI's decision to reverse its restructuring plans, the implications of this move, and the growing concerns among its critics. Featuring an in-depth conversation with Paige Headley, a former OpenAI employee and leader of the organization Not for Private Gain, the episode provides a comprehensive analysis of the tensions between profit motives and OpenAI's original charitable mission.

---

1. Introduction to OpenAI's Restructuring Debate

Stephen Overle begins by outlining the recent turmoil within OpenAI. Originally established as a nonprofit in 2015, OpenAI has been navigating the challenging transition toward a for-profit model under CEO Sam Altman. This shift aimed to attract significant investments from prominent backers like Elon Musk and Microsoft. However, recent developments have seen OpenAI partially reverse this course by planning to become a Public Benefit Corporation (PBC) while maintaining its nonprofit roots.

“The maker behind ChatGPT has been part of a nonprofit since it was founded back in 2015. But lately CEO Sam Altman has been trying to make it just a for-profit company.” [00:33]

2. The Reversal: Becoming a Public Benefit Corporation

OpenAI's latest strategy involves restructuring as a Public Benefit Corporation, allowing it to pursue both profit and its original mission to ensure that artificial general intelligence (AGI) benefits humanity. Sam Altman expressed agreement with this decision, asserting that both the nonprofit and PBC entities would share the same mission.

“It is now planning to become a public benefit corporation but remain part of its nonprofit.” [00:33]

3. Criticism from OpenAI's Former Supporters

Despite the apparent compromise, critics like Paige Headley remain unconvinced. Headley, representing Not for Private Gain, argues that the restructuring does not sufficiently protect OpenAI's charitable mission. She contends that the new structure may still prioritize commercial interests over public welfare.

“Based upon what I've seen publicly, this does not seem at all like a victory.” [02:06]

4. Governance and Control: Nonprofit vs. Public Benefit Corporation

A significant part of the discussion centers on governance. Under the new structure, the nonprofit would retain some control over the PBC, potentially through special voting rights. However, Headley emphasizes that this may not equate to true economic control, raising concerns about whether the nonprofit can effectively oversee the PBC's actions.

“They can only require the entity to act in accordance with its fiduciary duties.” [04:38]

“In the current structure, all commercial interests are subordinate to the charitable mission of the nonprofit.” [03:36]

5. Influence of Major Investors and the Role of Elon Musk

The episode explores whether influential figures like Elon Musk have significantly impacted OpenAI's restructuring plans. Headley suggests that the changes are more in response to legal pressures from the attorneys general of Delaware and California rather than Musk's activism.

“Musk is asking for a number of things. Their response does not seem particularly tailored to Musk's requests.” [10:03]

6. Balancing Profit Motives with Public Mission

Headley highlights the inherent tension between OpenAI's need to innovate and remain competitive, especially against global competitors like China, and its obligation to prioritize public welfare. She argues that profit motives can often conflict with the organization's charitable mission, particularly as OpenAI's valuation and influence grow.

“There is a tension between the interests of shareholders, the interests of the company and the interests of the public, and there always will be.” [16:05]

7. The Role of Government Regulation

The episode touches on the diminishing appetite for AI regulation in the U.S. government, contrasting it with the rising stakes of OpenAI's internal governance decisions. Headley expresses concern that without robust regulatory frameworks, companies like OpenAI may prioritize profits over safety and ethical considerations.

“If a company is behaving very irresponsibly, just looking out for its bottom line without regard to the collateral costs... it's quite scary.” [19:12]

8. Future Implications and Continued Advocacy

Despite the partial reversal of its restructuring plans, Headley remains skeptical about the sufficiency of OpenAI's new structure in safeguarding its original mission. She calls for more concrete commitments to prioritize public welfare over commercial gain and indicates that her organization will continue to advocate for stronger governance measures.

“I think that would not happen by default and I think if OpenAI were currently planning to do that, they would have said that in their announcement.” [19:58]

9. Conclusion: An Ongoing Struggle for OpenAI's Mission

Stephen Overle concludes the episode by acknowledging that OpenAI's restructuring is far from a settled matter. The tensions between profit and public good are likely to persist, necessitating continued scrutiny from critics, regulators, and stakeholders to ensure that OpenAI remains true to its foundational mission.

“It sounds like your fight continues on unfortunately. So Paige, thanks for being here.” [20:32]

---

Key Takeaways:

• OpenAI's Restructuring: Transitioning to a Public Benefit Corporation while maintaining its nonprofit origins.
• Critics' Concerns: Skepticism about the effectiveness of the new structure in prioritizing the public mission over commercial interests.
• Governance Challenges: Uncertainty about the true extent of nonprofit control over the PBC and the implications for OpenAI's strategic decisions.
• Influence of Investors: Legal pressures from state attorneys general appear to be a significant factor in the restructuring, rather than solely investor activism.
• Regulatory Landscape: The lack of stringent AI regulation in the U.S. heightens the importance of internal governance structures within AI companies.
• Ongoing Advocacy: Organizations like Not for Private Gain are likely to continue pushing for stronger safeguards to ensure that AI development remains aligned with public interests.

This episode provides a critical lens on OpenAI's evolving structure, highlighting the complexities and challenges of balancing innovation, commercial success, and ethical responsibility in the rapidly advancing field of artificial intelligence.

Transcript

[Uber Advertisement]

You know that feeling when someone shows up for you just when you need it most? That's what Uber is all about. Not just a ride or dinner at your door. It's how Uber helps you show up for the moments that matter. Because showing up can turn a tough day around or make a good one even better. Whatever it is, big or small, Uber is on the way. So you can be on yours. Uber on our way.

[Stephen Overle]

Hey, welcome back to Politico Tech. I'm your host Stephen Overle. The saga over control of OpenAI took yet another turn this week. The maker behind ChatGPT has been part of a nonprofit since it was founded back in 2015. But lately CEO Sam Altman has been trying to make it just a for profit company. And in the process, he's pitted early backers like Elon Musk against more recent investors like Microsoft. Well, this week OpenAI reversed course, sort of. It is now planning to become a public benefit corporation but remain part of its nonprofit. And speaking to reporters, Altman said he agreed with that decision and that the two entities will share the same mission. Altman could face questions about that new plan today when he appears before the Senate Commerce Committee. And that's because critics of the for profit switch still have questions, like Paige headley, a former OpenAI employee. He leads an organization called not for Private Gain that opposes the change and wrote an open letter calling on the attorneys general of Delaware in California to fight it. So when I called Page up to talk, I expected to hear some excitement that his group got what it wanted, gloating even. And that's not what I got.

[Paige Headley]

Based upon what I've seen publicly, this does not seem at all like a victory.

[Stephen Overle]

On the show today, Paige tells me why this news isn't necessarily cause for a victory lap. Here's our conversation.

[Unknown]

Foreign.

[Stephen Overle]

Welcome to Politico Tech.

[Paige Headley]

Thank you for having me.

[Stephen Overle]

So OpenAI decided this week that it would remain under the control of a nonprofit rather than spinning off into a for profit entity. I know this is something that you advocated for sum up for me why you see this as a good thing.

[Paige Headley]

I don't necessarily see it as a good thing. I think it's worth diving into what this might mean.

[Unknown]

The details are unclear.

[Paige Headley]

What is the same is the current for profit entity, which is the llc.

[Unknown]

Is still going to convert into a public benefit corporation.

[Paige Headley]

That was the plan before, that's the plan now. What has changed is the nonprofit entity will have some sort of control over that pbc.

[Unknown]

It doesn't look like it's going to be economic control. And so, based upon reporting from a few months back, I think the most.

[Paige Headley]

Likely candidate is voting control. They would give the nonprofit special voting rights.

[Unknown]

And so then the question is, what.

[Paige Headley]

Is the import of the nonprofit having special voting control? I think the answer is not necessarily that much, and we can talk about.

[Unknown]

Why in more detail.

[Paige Headley]

But the short answer is, right now.

[Unknown]

In the current structure, all commercial interests.

[Paige Headley]

Are legally subordinate to the charitable mission of the nonprofit. Everything OpenAI does must be consistent with it. If it's not, the board has the.

[Unknown]

Authority to step in.

[Paige Headley]

If the board does not step in, the attorneys general can step in and they have recourse in the courts. Under the current plan, the Public Benefit.

[Unknown]

Corporation would not have that same obligation, would not be required to put the charitable mission above commercial interests.

[Paige Headley]

So no matter who is on the.

[Unknown]

Board, no matter who has voting rights, that is a significant difference.

[Stephen Overle]

Got it. So you're saying the motivations here will still change. OpenAI will still be kind of commercially corporate driven, even though this nonprofit will retain some type of control, at least in terms of how the organization is structured.

[Paige Headley]

Yes, the nonprofit would have presumably voting control, but that does not mean they would be able to force the commercial.

[Unknown]

Entity to act in a certain way.

[Paige Headley]

They can only require the entity to.

[Unknown]

Act in accordance with its fiduciary duties.

[Paige Headley]

And so what matters here above all.

[Unknown]

Else is what are the fiduciary duties of the commercial entity, the Public Benefit Corporation? And those are very different than the duties of a charity.

[Stephen Overle]

Got it. And so I was going to ask you this question, which is, how much of a difference does this structure make? Particular, particularly if the leadership remains the same. You know, CEO Sam Altman is still going to be in charge. Any organization is ultimately just a reflection of the people at the top.

[Paige Headley]

That's right.

[Unknown]

The people at the very top right.

[Paige Headley]

Now, in principle, are the members of the nonprofit board. Now, whether they are empowered to discharge.

[Unknown]

Their fiduciary duties is another question. That's a hard job. OpenAI is an organization with thousands of people and a $300 billion valuation.

[Paige Headley]

Meaningfully overseeing what OpenAI does is difficult. That said, the nonprofit board in the new structure would not have the same.

[Unknown]

Role that it has now.

[Paige Headley]

Right now, it has complete management control.

[Unknown]

It has sole control over everything OpenAI does.

[Stephen Overle]

So when Sam Altman told reporters earlier this week that he viewed this decision as a good thing, that he agreed with it, do you believe him?

[Paige Headley]

I don't know what Sam had in mind when he said that.

[Unknown]

I certainly view it as a good.

[Paige Headley]

Sign that OpenAI's board is listening to the concerns that people have and trying to find a better way forward. I don't think what we've seen yet is what is needed to preserve their duties to their mission. But it's a sign that they're open.

[Unknown]

Minded, that they're still trying to figure this out.

[Paige Headley]

I don't know what Sam wants in all of this, so I can't speak.

[Unknown]

To specifically what what he found promising.

[Paige Headley]

About the current plan.

[Stephen Overle]

Sam Altman is actually testifying today before the Senate Commerce Committee in this hearing that's really all about US competitiveness around AI, particularly vis a vis China. And that was kind of a big reason for this proposed change to begin with, right? This idea that it would make OpenAI, and by extension the US more innovative, more competitive. Why weren't you sold on that argument?

[Unknown]

There are a few reasons.

[Paige Headley]

Number one, OpenAI is not America's only.

[Unknown]

Organization at the frontier of AI.

[Paige Headley]

There are a number all the companies in China building AI are I think, by all accounts behind. And so while you might think it's important that American companies stay ahead of right now, America seems pretty securely at the front and that gives them the freedom to pursue this goal responsibly.

[Unknown]

They can take the time to ensure they are adequately testing these models before.

[Paige Headley]

They'Re deployed, not endangering the public while.

[Unknown]

Still maintaining a lead.

[Paige Headley]

If OpenAI were handicapped by its commitments to the public, I think America would still be okay if we thought that moving quickly to stay ahead of China.

[Unknown]

Was of paramount importance.

[Paige Headley]

OpenAI has other options available to it. The stop and assist commitment in the OpenAI charter, the idea of that commitment was in a neck and neck race to build a very powerful technology.

[Unknown]

Organizations or countries might have powerful incentives.

[Paige Headley]

To move very quickly, even if doing so was dangerous and that could endanger the public. And so that's why OpenAI committed to not doing that, provided one of the other competitors shared its goals. So in the context of China, if OpenAI really believed that it was important to make American companies move faster, it has an option available to it that's not available as easily to other for profit companies.

[Unknown]

It can give all of its resources.

[Paige Headley]

Its staff, its ip, its computer to another American company that's trying to do.

[Unknown]

The same thing it's trying to do. And that would presumably help American competitiveness.

[Paige Headley]

Far more than removing its obligations to the public as a nonprofit.

[Stephen Overle]

It would torpedo though OpenAI as a business entity or with its own business interests.

[Paige Headley]

Well, two responses to that one if we're thinking about the interests of America relative to China, that doesn't matter. OpenAI is just an American company. And if what we're talking about is advancing American interests, then we should not care whether OpenAI is no longer a going concern.

[Unknown]

Second, OpenAI is not a commercial company right now.

[Paige Headley]

It is a nonprofit.

[Unknown]

The only reason it's allowed to try to raise money is in order to achieve its charitable goals. And so if trying to raise money.

[Paige Headley]

Conflicts with its charitable goals, it must pursue the first thing differently.

[Stephen Overle]

Right. Well, and that's one of the criticisms of this proposed change, is this kind of idea of a bait and switch, that it was started and raised all this money with one mission in mind, you know, as one type of organization. Now it's trying to pivot that. It has sort of had a whiff of commercial success with things like ChatGPT and other programs it's created. You know, I have to ask, how much of this outcome, you know, OpenAI, sort of walking things back or at least pursuing a different plan, do you view as the direct result of Elon Musk? You know, he filed a lawsuit. He's really been agitating here. How much of this is because of him?

[Paige Headley]

I'm not sure. Musk is asking for a number of things. Their response does not seem particularly tailored to Musk's requests.

[Unknown]

I think it might be more tailored.

[Paige Headley]

To the requests and the concerns of the attorneys general of Delaware and California. And we don't have much insight into.

[Unknown]

What those concerns are or what conversations they're having behind the scenes.

[Paige Headley]

But I view the current update as a positive sign that the attorneys general are taking this issue very seriously and asking the right questions.

[Unknown]

Summer's here, and Nordstrom has everything you need for your best dress season ever, from beach days and weddings to to weekend getaways in your everyday wardrobe. Discover stylish options under $100 from tons of your favorite brands like Mango Skims, Princess Polly and madewell. It's easy, too, with free shipping and free returns in store order pickup and more. Shop today in stores [e-mail] or download the Nordstrom app.

[Stephen Overle]

You know a bit about your background. You worked at OpenAI. You were a policy advisor and ethics Advisor there in 2017, 2018. I'm curious for you, why get involved in this fight now? Like, what brought you back into this?

[Unknown]

I had little interest in being involved in this fight.

[Paige Headley]

I was watching this happen from the sidelines.

[Unknown]

I was a bit saddened by it.

[Paige Headley]

A bit unsurprised by it, and I expected there to be a lot of pushback from other people in the industry, from the media, from the public, and was not seeing the pushback that I expected. And so at some point I thought it made sense for me to be a bit more involved and I did that. Gradually I became more immersed in the issues and then ultimately decided with some of the people that signed the letter alongside me to say something publicly to the attorneys general.

[Stephen Overle]

What's the right question that journalists like myself should be asking now? I mean, I know there's many sort of outstanding questions about exactly what this new structure would look like, whether folks like, you know, Microsoft and other investors will sign off on it. What do you sort of see as the big question everyone should be asking OpenAI, including Congress today, when Sam Altman is there?

[Paige Headley]

I think the correct question is would this advance OpenAI's charitable mission, that is to ensure artificial and general intelligence is developed safely and for the benefit of humanity and not for private gain?

[Unknown]

I think everything should be viewed through that lens.

[Paige Headley]

Does this restructuring advance or undermine that mission?

[Stephen Overle]

What would it take to ease your concerns and those who kind of share your criticisms of what this transition would mean for OpenAI and for its mission?

[Paige Headley]

I think the most important thing is.

[Unknown]

If you look at the 2019 blog.

[Paige Headley]

Post from OpenAI that describes their restructuring.

[Unknown]

Into the current hybrid capped profit model, they excerpt many passages from the operating agreement of the new for profit entity. They state that all commercial interests are subordinate to the charitable mission and specifically the OpenAI charter. The OpenAI charter is a very important document. It is how OpenAI has operationalized its charitable mission.

[Paige Headley]

That's the document that includes things like the commitment to stop competing and assist a mission aligned company close to building artificial general intelligence. I would like to see a similar contractual commitment from the PBC that as they do today, all investors and all.

[Unknown]

Employees must sign, making it abundantly clear that the mission comes first.

[Stephen Overle]

You know, anytime I talk to a current or former employee, I'm always curious to ask whether you saw things or have insight into things that you don't think the public at large does. And I recognize that OpenAI in many ways is a different company now than it was in a different organization now than it was back when you worked there. But I'm curious what you saw that maybe the public doesn't get to see.

[Paige Headley]

I think absolutely. And I think that's one of the reasons I was motivated to get involved. When I worked at OpenAI, there was very much a discussion at the company about what OpenAI was these were early days. It was founded to be different, to be a response to profit driven AGI development.

[Unknown]

But what does that actually mean? What would make it different and what would make it robustly different going forward?

[Paige Headley]

And there are of course, different factions at the organization. Some people who were idealistic true believers in the mission, some people who just want, wanted to make cool tech products, and there's nothing wrong with that. And some people had more prosaic commercial or competitive motivations. And those discussions led to the OpenAI charter, which is right now basically OpenAI's constitution. And the idea was to codify, to write down what it is about the.

[Unknown]

Organization that makes it different, to articulate.

[Paige Headley]

This delicate but important balance between the need to raise capital, to be the frontier, while nonetheless remaining true to the charitable mission. And you can think about these as OpenAI's attempt to tie itself to the mast.

[Unknown]

So no matter how powerful profit motives.

[Paige Headley]

Became, and we always knew they'd become.

[Unknown]

Very powerful if OpenAI were successful, if it came close to building AGI, which.

[Paige Headley]

It is by all accounts right now, the amount of money, the amount of.

[Unknown]

Power at stake, would very predictably be extraordinary.

[Paige Headley]

So there had to be robust safeguards that could not be removed to ensure the organization, even during that very tempting.

[Unknown]

Phase, remained true to its mission.

[Paige Headley]

And I think what we're seeing right now is what we were worried about, that when the prize is close enough and big enough, those safeguards are chafing.

[Stephen Overle]

Which it's so fascinating to hear that because it's one attention kind of society wide in the conversation around AI, you know, talking about the balance of safety versus innovation, however you kind of want to frame it, but. But this idea about sort of chasing unfettered AI versus regulating it, it's also a tension that we've seen within OpenAI before. For instance, when Sam Altman was fired and brought back, all in the course of sort of a week. I wonder if there's a resolution to that tension or if that is something that is just going to exist as long as OpenAI exists.

[Paige Headley]

I think that tension is unavoidable. I think that the basic idea behind OpenAI's founding and the governance safeguards that our letter focuses on and that we think are at stake here is that there is a tension between the interests of shareholders, the interests of the company and the interests of the public, and there always will be. If there weren't, you wouldn't have needed the nonprofit in the first place.

[Unknown]

The interests will often be aligned.

[Paige Headley]

I'm a big fan of the Free market. I think for profit companies are great. I think most companies should be for profit.

[Unknown]

And it drives innovation, it drives prices down.

[Paige Headley]

You get cool products, but in some cases, the interest of the public and of the company's investors will be different. And for a technology as potentially transformative.

[Unknown]

As AGI, those differences could be significant.

[Paige Headley]

And so I think that tension is inherent and just needs to be managed.

[Stephen Overle]

You know, we have a lot of listeners in government and on the Hill, obviously, for a Politico tech podcast. The US Government has really backed off this idea of regulating AI. It was very much sort of a hot part of the discussion a year or two ago. But especially under the Trump administration, that appetite seems to have sort of left Washington, I guess. How does that change the stakes in this whole open AI battle, considering that AI companies may be left to largely regulate themselves?

[Unknown]

Yeah, I think that significantly increases the stakes here.

[Paige Headley]

And you can think about the floor of what's possible if a company is.

[Unknown]

Behaving very irresponsibly, just looking out for its bottom line without regard to the collateral costs.

[Paige Headley]

And to cite an example, right now, all frontier AI companies have committed, voluntarily committed, to testing their models for certain risks before they are made available to the public, such as the ability for regular people like us to create biological weapons that could kill much of the world.

[Unknown]

Seems pretty important to do.

[Paige Headley]

These tests are time consuming, they're difficult. And we have seen reports that in some cases, OpenAI has rushed through them such that the people who are running them have complained that they weren't done properly in order to meet something like.

[Unknown]

A product launch date.

[Paige Headley]

There is a nearby world in which there were some very basic targeted regulations that would require certain tests before a model is released. If we lived in that world, then the worst thing that OpenAI could do, consistent with the law, if it had no nonprofit oversight, would be less scary. But in the world that we're in.

[Unknown]

Right now, it's quite scary. So it increases the stakes of this entire issue.

[Stephen Overle]

You know, it's funny, when we started this conversation, I was kind of expecting it to be a bit of a victory lap for, like, the AI safety movement and, you know, the civics folks who have just been kind of agitating and pointing to some of these risks. I don't get the sense that you see this as a huge win, though.

[Paige Headley]

I don't see it as a win at all at the moment, with the caveat that, you know, there are a.

[Unknown]

Lot of details that have not been.

[Paige Headley]

Shared, and I might be pleasantly surprised. It might be that the Public Benefit.

[Unknown]

Corporation will be bespoke and will have.

[Paige Headley]

Commitments in its articles along the lines I described. I think that would not happen by default and I think if OpenAI were currently planning to do that, they would have said that in their announcement. So based upon what I've seen publicly, this does not seem at all like.

[Unknown]

A victory other than as I mentioned.

[Paige Headley]

It means that people are listening and open minded and there's still time to.

[Unknown]

Reach a better outcome for the mission.

[Stephen Overle]

Got it. So it sounds like your fight continues on unfortunately. So Paige, thanks for being here.

[Paige Headley]

Thank you very much. Really appreciate it.

[Stephen Overle]

That's all for today's Politico Tech. If you like Politico Tech, be sure to subscribe and recommend it to a friend or colleague. And for more tech news, subscribe to our newsletters, Digital Future Daily and Morning Tech. Music in our show comes from the mysterious Brake Master Cylinder. Our producer is Nirmal Molaikal. I'm Stephen Overle. See you back here next Thursday.