Former employees of OpenAI are urging attorneys general in California and Delaware to halt the company’s plans to shift its artificial intelligence control from a nonprofit trust to a for-profit corporation.
This shift is generating concerns about the consequences if OpenAI, the creator of ChatGPT, achieves its goal of developing AI that surpasses human capabilities while potentially abandoning its mission to prevent harm.
Page Hedley, a former policy and ethics advisor at OpenAI, expressed apprehension about the ownership and control of this powerful technology once developed.
Supported by distinguished figures, including three Nobel laureates, Hedley and nine other former OpenAI employees addressed a letter to the attorneys general of California and Delaware. They implore state officials to use their powers to protect OpenAI’s original charitable objectives and prevent its impending structural change. OpenAI’s headquarters are in San Francisco, and the company is incorporated in Delaware.
In response, OpenAI stated that any changes in its structure aim to broaden AI’s benefits to the public. The company intends to become a public benefit corporation, akin to other AI organizations such as Anthropic and Elon Musk’s xAI, while maintaining a nonprofit division.
“This structure will ensure that as the for-profit thrives and expands, so does the nonprofit, allowing us to fulfill our mission,” the company stated.
This letter marks the second this month addressed to state officials, following a petition from labor and nonprofit groups who seek to safeguard OpenAI’s significant charitable resources. Last fall, Delaware Attorney General Kathy Jennings expressed her willingness to scrutinize any transaction to ensure it aligns with public interest. Meanwhile, California Attorney General Rob Bonta’s office has requested more information but refrained from commenting further.
OpenAI’s founding members, including CEO Sam Altman and Musk, initially launched the organization as a nonprofit with the vision of safely developing artificial general intelligence (AGI) for humanity’s welfare. Nearly a decade later, the company boasts a market valuation of $300 billion and 400 million weekly users of ChatGPT.
Transitioning to a predominantly for-profit model poses significant challenges for OpenAI, including legal action from Musk, who accuses the enterprise and Altman of straying from the principles he invested in initially. While some who signed the letter back Musk’s lawsuit, others like Hedley remain skeptical due to Musk’s vested interests in his own rival AI company.
Notable contributors to the letter include Nobel-winning economists Oliver Hart and Joseph Stiglitz, as well as AI trailblazers Geoffrey Hinton and Stuart Russell. Hinton, in particular, expressed support for OpenAI’s original mission to ensure AGI benefits all of humanity rather than merely enriching investors.
Tensions around OpenAI’s objectives have existed for years, influencing significant decisions such as Musk’s 2018 departure and Altman’s brief exit in 2023. Hedley reflected on his experiences from 2017-18 and raised concerns about the organization’s accelerated product releases with limited safety assessments amid ChatGPT’s success.
“The risks associated with these choices will escalate as the technology becomes more advanced,” Hedley explains. “In the proposed new structure, there’s a heightened risk of hasty decisions without adequate oversight or restraint.”
Anish Tondwalkar, a former software engineer with OpenAI, emphasized the significance of a “stop-and-assist clause” within OpenAI’s nonprofit charter. This clause obligates OpenAI to intervene and support if another entity is close to achieving beyond-human AI.
Tondwalkar warned that transitioning to a for-profit model could erase these safeguards and OpenAI’s duty to serve the public interest, potentially overnight.
Nisan Stiennon, an ex-OpenAI engineer, succinctly stated the fear of the advanced technology going unregulated. “It’s to OpenAI’s credit that it’s governed by a nonprofit with responsibilities toward humanity, which shouldn’t be relinquished.”
The debate continues over the balance between profit and public responsibility in AI advancement, highlighting the complexities of ethical technology development.