OpenAI rewrites its founding principles for the first time since 2018, dropping AGI focus and embracing mass deployment
Sam Altman calls earlier caution about restricting powerful models a "misplaced worry" as the company shifts from research lab to infrastructure provider.
OpenAI has published its first major update to the principles that govern its mission since 2018, replacing an AGI-centric charter with five broad commitments that signal a decisive shift from cautious research laboratory to aggressive deployer of AI infrastructure at a planetary scale.
The 2018 charter mentioned artificial general intelligence twelve times and treated its development as the company's explicit north star.
The 2026 document mentions it twice.
Chief executive Sam Altman framed the change as a natural evolution of the iterative deployment strategy OpenAI developed after its initial nervousness about releasing GPT-2 in 2019, when the company withheld the full model over concerns it could be misused.
"We worried a lot in the early days about the risks of releasing models, and erred on the side of restricting access," Altman wrote.
"In retrospect, it was a misplaced worry."
The five new principles commit OpenAI to democratising access to AI, empowering users with greater control over their tools, building resilient safety systems, adapting corporate governance as capabilities advance, and delivering what the company calls "universal prosperity" through large-scale AI infrastructure.
AGI, Altman wrote, has a "ring of power" quality "that makes people do crazy things."
The only solution, he argued, is to "orient towards sharing the technology with people broadly, and for no one to have the ring."
The shift in tone is stark.
The 2018 charter explicitly pledged to stop competing and assist any "value-aligned, safety-conscious project" that came closer to building AGI, with a suggested trigger of a "better-than-even chance of success in the next two years."
The 2026 document makes no such commitment, instead acknowledging that OpenAI "is a much larger force in the world than it was a few years ago" and pledging transparency about when and how its principles might change.
Business Insider identified three meaningful departures from the original charter: the new principles imply OpenAI could prioritise its own interests over universal accessibility in certain circumstances; the collaborative tone towards rival laboratories has been replaced by a competitive stance; and the document explicitly acknowledges that future model capabilities may be restricted for safety reasons, a caveat absent from the original.
The timing is not incidental.
OpenAI is navigating its conversion from a capped-profit entity to a fully commercial structure, a process being challenged by Elon Musk in an Oakland courtroom this week, and is exploring an initial public offering that could value it at up to $1 trillion.
The principles update lands days after the company released GPT-5.5, which co-founder Greg Brockman described as "a new class of intelligence."
Related reading
- OpenAI publishes open-source teen safety tools for developers building AI apps
- OpenAI strengthens Sora video safety and consent controls
- DeepMind scientist argues no AI system will ever become conscious, calling the assumption a 'fundamental fall…
"What's really special about this model is how much more it can do with less guidance," Brockman said.
The company also quietly removed the word "safely" from its mission statement in its most recent IRS disclosure, a change first reported by Fortune in February, prompting concerns from former employees and safety researchers that commercial pressures are overriding the caution that defined OpenAI's earlier years.
The recap
- OpenAI updates mission to pursue broadly distributed AGI.
- The company published a 1,100-word document on its website.
- GPT-5.5 is available to ChatGPT Plus, Pro, Business and Enterprise users.
