Artificial, In General
In recent weeks, OpenAI quietly changed its “core values” list to include a focus on artificial general intelligence (AGI) that wasn’t explicitly listed there before.
As Semafor reports, the firm as recently as September 21 listed its values on a job openings page as “Audacious,” “Thoughtful,” “Unpretentious,” “Impact-driven,” “Collaborative” and “Growth-oriented,” per a snapshot of the page on the Internet Archive.
When one goes to the page now, though, they’ll see an entirely different list that includes “AGI focus” as its very first value. Its other new core values: “Intense and scrappy,” “Scale,” “Make something people love,” and “Team spirit.”
Sure, it’s all corporate blather. But you can’t help but wonder: if you can replace all your core values at the drop of a hat, were they really core values to begin with?
Whiplash!
And let’s zoom in on that number one core value, “AGI Focus,” because it’s a perfect example of how the company’s favorite terms can feel like works in progress.
In February, OpenAI’s inscrutable CEO Sam Altman wrote in a company blog post that AGI can broadly be defined as “systems that are generally smarter than humans,” but in a wide-ranging New York Magazine interview published last month, he’d downgraded the definition to AI that could serve as the “equivalent of a median human that you could hire as a co-worker.”
So which is it, then? Does OpenAI and its CEO think that AGI, its purported new core value, will be comprised of superhuman artificial intelligence, or is it an AI that’s just about as smart as the average person?
Shifting Tides
While we’ve reached out to OpenAI to clear up that pretty wide discrepancy in definition, the answer may not become clear anytime soon given that the company’s goals have, like its values, shifted over time as well.
Founded in 2015 by Altman, Elon Musk, and a handful of others who are by and large no longer affiliated, OpenAI was created as a nonprofit research lab that was meant, essentially, to build good AI to counter the bad. Though the firm still pays lip service to that original goal, its drift away from nonprofit AI do-gooders to a for-profit endeavor led to Musk’s exit in 2019, and that purpose-shifting appears to have bled into its self-descriptions as well.
So yeah, the new core values list feels like a bunch of fluff. It does, however, seem to show OpenAI’s hand when it comes to its single-minded focus looking forward.
“We are committed to building safe, beneficial AGI that will have a massive positive impact on humanity’s future,” the OpenAI job postings page now explains. “Anything that doesn’t help with that is out of scope.”
More on OpenAI: AI’s Electricity Use Is Spiking So Fast It’ll Soon Use as Much Power as an Entire Country