“A.I. doesn’t dream. It doesn’t feel. It doesn’t make art. It aggregates it. Our job as creators, as storytellers, and as publishers is to make people feel something real. That’s why we create, and that’s why we’re still here.”
—Jim Lee, the legendary artist and DC Comics co-publisher, explaining why his company won’t support A.I.-generated storytelling or artwork.

And now for the main event…

Over the weekend, Nathan Calvin, the general counsel for A.I. safety nonprofit Encode, accused OpenAI of engaging in legal intimidation tactics in a viral post on X. He detailed a nerve-racking experience, in early August, during which OpenAI served him two subpoenas: the first addressed to Encode, and the second addressed to him directly.

Jason Kwon, OpenAI’s chief strategy officer, downplayed the subpoenas as a routine fact-finding exercise amid the company’s ongoing battle with Elon Musk. After all, more than a year ago, Musk embarked on a legal campaign to prove that OpenAI had betrayed its founding mission by attempting to convert into a for-profit organization. (Musk was a founding member of OpenAI.) The subpoena, Kwon noted, merely sought to ascertain whether there was a financial connection between Musk and Encode, given that the nonprofit group had filed an amicus brief supporting some of Musk’s arguments. “For a safety policy organization to side with Elon (?), that raises legitimate questions about what is going on,” Kwon wrote on X.

The legal notice to Encode was just the latest in a flurry of subpoenas that OpenAI has issued to a series of nonprofits in recent months, after the company countersued Musk in April alleging corporate harassment. The subpoenaed entities include The Midas Project, Legal Advocates for Safe Science and Technology (LASST), and the Coalition for AI Nonprofit Integrity (CANI). In some ways, this activity is hardly a surprise: OpenAI’s last two funding rounds are at least partially contingent on the company successfully completing its restructuring into a public benefit corporation. Hundreds of billions of dollars are at stake—literally.

Calvin viewed the situation much differently, accusing OpenAI of having “used the pretext of their lawsuit against Elon Musk to intimidate their critics and imply that Elon is behind all of them.” Whatever Altman’s intentions, there’s no question that the subpoenas have been far-reaching. The legal documents that were serviced to Encode requested information that went far beyond the organization’s dealings with Musk. It asked for “all documents and communications concerning the governance or organizational structure of OpenAI,” and all documents “concerning SB 53 or its potential impact on OpenAI.” (SB 53, California’s recently enacted A.I.-focused legislation, was co-sponsored by Encode.)

On Sunday, I connected with Calvin, who told me that Encode submitted objections and responses to the subpoena, contending that the information OpenAI requested was either irrelevant to its case or didn’t exist. He said he never heard back. OpenAI “could file a motion to compel, but they never did that, and they never responded at all,” he said. On Monday, Encode’s founder and president, Sneha Revanur, said that “for us, this was taxing.” She added that “the legal fees could’ve been crushing.” Encode’s team consists of just three people.

Meanwhile, Calvin’s thread prompted a surprising response from Joshua Achiam, OpenAI’s head of mission alignment, pushing back against the subpoenas—a move that, he said, is “possibly a risk to my whole career.”  “We can’t be doing things that make us into a frightening power instead of a virtuous one,” he wrote on X. “The dangerously incorrect use of power is the result of many small choices that are all borderline but get no pushback; without someone speaking up once in a while it can get worse. So, this is my pushback.” When I asked for comment, an OpenAI spokesperson directed me to Kwon’s original thread.

Given their engagement with Musk’s lawsuit, Revanur acknowledged that some scrutiny of Encode was “reasonable.” However, it’s also true that OpenAI sent subpoenas to several organizations nowhere near the lawsuit, or with any obvious public connection with Musk. In May, LASST, a two-person organization founded last year by attorney Tyler Whitmer, was hit with a subpoena. Whitmer told me he viewed the measure as a good-faith misunderstanding. He said that his company had never taken money from Musk, doesn’t plan to, and that LASST is even working on public advocacy efforts against xAI.

The subpoena, he said, seemed to be a response to an open letter LASST published in partnership with Encode opposing OpenAI’s restructuring plans. It asked for documents and communications concerning California’s infamous SB 1047, the state’s failed precursor to SB 53. “They just discount the idea that a lot of us might genuinely believe that OpenAI continuing to be a nonprofit and continuing to control the commercial enterprise is very, very important for the future of humanity,” said Whitmer, adding that LASST had nothing to do with the bill.

Meanwhile, around the time Calvin was served, Tyler Johnston, the founder of nonprofit A.I. watchdog The Midas Project, received a double subpoena. He described a similarly broad document, seeking detailed information about his organization’s funding sources beyond Musk, including anyone associated with him, as well as with Mark Zuckerberg, or anyone related to Meta. It also sought all documents and communications concerning OpenAI’s governance structure and planned changes to that governance structure.

At the time of the subpoena, The Midas Project had become well known for two things: an open letter, with around 10,000 signatures, calling for transparency around several points regarding OpenAI’s proposed restructuring; and The OpenAI Files, a 14,000-word research document detailing The Midas Project’s concerns with the proposed restructuring. Both were referenced in the subpoena, which Johnston said encompassed “every conversation I had with a journalist; every congressional office that we talked to; every former employee we talked to.” After responding to the subpoena, Johnston told me, he never heard back from OpenAI.

For his part, Kwon defended the broad requests as being “standard language,” and said that The Midas Project’s focus on OpenAI’s restructuring raised “transparency questions about who was funding them and whether there was any coordination.” Johnston said he would have happily revealed that his company was not funded by Musk, either in court or over the phone. The breadth of the request, he said, seemed to suggest an attempt at intimidation. “Maybe they were trying to slow us down by burying us in paperwork in these critical months before this restructuring is finalized,” Johnston told me. He added that The Midas Project has been unable to obtain media liability insurance, and that the denials cited OpenAI’s subpoena as evidence of too much risk.

I reached out to a number of other safety-related organizations, including the AI Governance and Safety Institute, the AI and Democracy Foundation, PauseAI, the Center for Existential Safety, and Brown University’s AI Safety Team, among others, to check whether any of them had received subpoenas as well. None had been served, although some entities, like the Center for Existential Safety, have been expecting to receive one for years. However, I’ve learned that Eko.org, a nonprofit that published a petition seeking to block OpenAI’s proposed transition, was served with a similar subpoena—despite the fact that the group had already told OpenAI it was not funded by Musk.

That’s all for today. I’ll see you Thursday.

Ian

Need help? Review our FAQ page or contact us for assistance. For brand partnerships, email ads@puck.news.

You received this email because you signed up to receive emails from Puck, or as part of your Puck account associated with {{customer.email}}. To stop receiving this newsletter and/or manage all your email preferences, click here.

Puck is published by Heat Media LLC. 107 Greenwich St, New York, NY 10006