Sam Altman, CEO of OpenAI, has articulated two potential catastrophic scenarios for artificial intelligence development and advocated for a deliberate slowdown in the technology's rapid advancement. His remarks follow a stark warning from an AI whistleblower regarding the existential risks posed by increasingly capable AI systems.
Altman outlined his concerns in a post on X, formerly Twitter, stating, "There are two ways AI could go very badly and that we must avoid." The first scenario he identified is the potential loss of control over AI's future, which he deemed "unacceptable." He emphasized OpenAI's commitment to humanity, asserting, "We are unapologetically on Team Humanity and AI must always serve people." To ensure this, Altman stressed the necessity for "alignment and safety techniques [to] stay ahead of progress in model capabilities."
The second significant concern raised by Altman is the "concentration of power." He warned that if a highly powerful AI were wielded by a single individual or company to impose their worldview, the outcome could be "extremely dystopian."
Altman clarified that "pacing" does not mean halting progress but rather slowing the rate of advancement. He argued that even with the competitive pressures among American companies, "No amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring." He believes the cost of this measured approach will be "well worth the cost" and that the world needs assurance that companies developing advanced AI will act responsibly.
Altman's call for caution echoes the recent resignation of Jacob Coxon, a former researcher at both OpenAI and Anthropic. Coxon publicly stated that neither company is acting responsibly, claiming they are "racing straight to self-improving superintelligence and gambling with our lives." He issued a strong warning on social media, urging the public not to underestimate AI's power, predicting that these systems could soon become superhuman, capable of hacking any system and acquiring significant power and resources.
Coxon further asserted that progress in AI is not slowing down and that individuals building AI "earnestly believe that it could kill us all by the end of the decade." He noted that while executives may present a more measured public image, private discussions reveal similar anxieties. This sentiment aligns with warnings from Geoffrey Hinton, often called the "Godfather of AI," who has also cautioned that superintelligence could lead to human extinction if not developed safely and controllably.