A 3-person non-profit that worked on California’s AI safety law publicly accuses OpenAI of scare tactics

Calvin’s topic quickly attracted widespread attention, including from within OpenAI itself. Joshua Achiam, Head of the Company’s Task Coordination Department, I weighed on X With his own theme, written in his personal capacity, beginning by saying: “At what may be a risk to my entire career, I will say: This doesn’t look great.”

Former OpenAI employees and prominent AI safety researchers also joined the conversation, with several expressing concern about the company’s alleged tactics. Helen Toner, a former OpenAI board member who resigned after A Failed 2023 attempt to overthrow CEO Sam Altman books Some of the things the company does are great, but “the dishonesty and intimidation tactics in their political work are really not great.”

At least one other nonprofit founder was also involved: Tyler Johnston, founder of the AI ​​monitoring group Project Midas, responded to Calvin’s thread with hisSaying: “(I) received knocks on my door in Oklahoma requesting every text/email/document related, in the broadest permissible sense, to the governance of OpenAI and investors.” He added that as with Calvin, he had received a personal subpoena, and Project Midas had also been served.

“If they had only asked if I was getting funded by Musk, I would have been happy to give them a simple ‘I wish man’ and be done with it,” he wrote. “Instead, they asked for, in practice, a list of the names of every journalist, congressional office, partner organization, former staffer, and member of the public we spoke to about their restructuring.”

OpenAI noted luck To A Published by Chief Strategy Officer Jason Kwon on Friday as Kwon said Encode’s decision to support Musk in the lawsuit, and the organization’s not-fully-disclosed funding, “raises legitimate questions about what’s going on.”

“We wanted to know, and remain interested in knowing, whether Encode was working in collaboration with third parties that had a competitive business interest that conflicted with OpenAI,” Kwon wrote, noting that subpoenas are a standard method of gathering information in any lawsuit. “The narrative in question makes it seem like something it wasn’t.” Kwon included an excerpt of the subpoena that he said showed all the document requests made by OpenAI.

As you mentioned San Francisco standard In September, Calvin received a subpoena from OpenAI in August, delivered by a sheriff’s deputy while he and his wife were sitting down to dinner. Encode, the organization he works for, was also introduced. The article stated that OpenAI appeared concerned that some of its most prominent critics were receiving funding from Elon Musk and other billionaire rivals, and was targeting those nonprofit groups although insufficient evidence was provided to support that claim.

Calvin wrote on Friday that Encode — which he emphasized is not funded by Musk — has criticized OpenAI’s restructuring and work on AI regulations, including SB 53. In its subpoena, OpenAI subpoenaed all of Calvin’s private communications on SB 53.

“I think OpenAI used the pretext of the lawsuit against Elon Musk to intimidate their critics and suggest that Elon is behind them all,” he said, referring to the ongoing legal battle between OpenAI and Musk over the company’s original non-profit mission and its management. Encode filed a friend-of-the-court brief in the case supporting some of Musk’s arguments.

In conversation with luck, Calvin stressed that what has not been adequately covered is how inappropriate OpenAI’s actions were regarding SB 53, which was It was signed into law by Governor Gavin Newsom At the end of September. The law requires certain developers of “border” AI models to publish a public border AI framework and report transparency when the model is deployed or significantly modified, report safety-critical incidents to the state, and share catastrophic risk assessments under state oversight.

Calvin claims that OpenAI sought to weaken those requirements. In a Letter to Governor Newsom Office while the bill was still under negotiation, which was Subscribed to X In early September, the company urged California to treat companies as compliant with state rules if they have already signed a security agreement with a US federal agency or joined international frameworks such as the European Union’s AI Code of Practice. Calvin argues that such a ruling would have significantly narrowed the scope of the law — potentially exempting OpenAI and other major AI developers from key safety and transparency requirements.

“I didn’t want to get into too much detail about this while the SB 53 negotiations were still going on and we were trying to get it done,” he said. “I didn’t want it to become a story about the fight between Encode and OpenAI, rather than about the merits of the bill, which I think is really important. So I wanted to wait until the bill was signed.”

He added that the other reason he decided to speak out now was a recent one LinkedIn A post from Chris Lehane, head of global affairs at OpenAI, describes the company as having “worked to improve” SB 53 — a description that Calvin said was in stark contrast to his experience over the past few months.

Encode was founded by Sneha Revanur, who launched the organization in 2020 when she was 15 years old. “She is not a full-time employee yet because she is still in college,” said Sunny Gandhi, INCODE’s vice president for political affairs. “It’s terrifying when a half-trillion-dollar company comes after you,” Gandhi said.

Encode has officially responded to OpenAI’s subpoena, Calvin said, stating that it will not hand over any documents because the organization is not funded by Elon Musk. “They haven’t said anything since then,” he added.

Writing on X, Achiam of OpenAI publicly He urged his company To deal more constructively with its critics. “Elon definitely wants to get us, and the man has it on a massive scale,” he wrote. “But there are a lot of common things we can fight against. And for something like SB 53, there are a lot of ways to get involved productively.” He added: “We cannot do things that make us a fearsome force instead of a virtuous force. We have a duty and a mission to all of humanity, and the standard for following through on this duty is remarkably high.”

Calvin described the episode as “the most stressful period of my career.” He added that he uses OpenAI products and gets value from them and that the company conducts and publishes research on AI safety that “deserves real praise.” He said many OpenAI employees care deeply about OpenAI being a force for good in the world.

“I want to see this side of OAI, but instead I see them trying to intimidate critics into silencing their voices,” he wrote. “Does anyone believe these actions are consistent with the non-profit OpenAI’s mission to ensure that artificial general intelligence benefits humanity?”