What you need to know
- OpenAI has lost almost half of its super alignment team.
- A former OpenAI researcher attributes the mass exodus of employees from the firm to its focus on shiny products as safety processes take a backseat.
- The research claims OpenAI could be on the brink of achieving AGI (artificial general intelligence), but it cannot handle what it entails.
Over the past few months, many OpenAI staffers have departed from the tech startup for various reasons. Recently, OpenAI co-founder Greg Brockman announced that he’d be going on sabbatical leave through the end of the year while researcher John Schulman announced his departure from the firm to join Anthropic to focus on AI alignment.
The mass exodus of high-profile executives from the AI firm began shortly after Sam Altman’s bizarre firing and reinstatement as CEO by the board of directors, including former OpenAI super alignment lead Jan Leike, who indicated that he left the company after falling into multiple disagreements with top officials at the company over safety, adversarial robustness, and more. Leike also noted that safety procedures took a backseat, giving precedent to shiny products.
While touching base with Fortune, Daniel Kokotajlo, who also worked as a researcher at OpenAI until early 2023, indicated that more than half of OpenAI’s super alignment team has already departed from the company. “It’s not been like a coordinated thing. I think it’s just people sort of individually giving up,” added Kokotajlo.
As you may know, OpenAI’s founding mission was to “ensure that artificial general intelligence benefits all of humanity.” However, the firm has veered off this goal and is now a regular for-profit company.
Elon Musk has openly criticized OpenAI for abandoning its founding mission, referring to it as a “stark betrayal.” The billionaire filed a lawsuit against the firm and Sam Altman over the same issue, but abruptly withdrew it earlier this year. Interestingly, the Tesla CEO launched another complaint against the AI firm — citing involvement in racketeering activities. According to Musk’s lawyers, “The previous suit lacked teeth.”
Is OpenAI AGI-ready?
It’s no secret that OpenAI is working toward hitting the AGI benchmark, however, there’s a rising concern among users about its implications for humanity. According to an AI researcher, there’s a 99.9% probability AI will end humanity, and the only way to stop this outcome is not to build AI in the first place.
Although OpenAI has since formed a new safety team led by CEO Sam Altman to ensure the company’s technological advances meet critical safety and security standards, it’s seemingly more focused on product development and the commercial side of business.
Interestingly, months after disbanding its safety alignment team, OpenAI reportedly rushed through the GPT-40 launch and even sent out invitations for the event before testing began. The company admitted the safety and alignment team was under pressure and left with little time for testing.
To this end, it’s impossible to identify the root cause for many executives and staffers who’ve left the company in the past few years, with some even starting rival firms focused on making superintelligence safe. Kokotajlo speculates the mass departure is directly related to OpenAI being on the precipice of hitting the AGI benchmark, but it lacks the knowledge, regulations, and tools “to handle all that it entails.”
Elsewhere, OpenAI recently opposed a proposed AI bill by Senator Scott Wiener designed to install safety protocols to prevent the technology from veering off the guardrails, citing a need for federal legislation.