OpenAI has released a new framework that is designed to help the company sustain itself beyond its co-founder and CEO Sam Altman.
What Happened: OpenAI's new "Preparedness Framework" has been released, empowering its board of directors to veto decisions taken by the company's leadership, which Altman currently leads.
While OpenAI's leadership team continues to have the power to make decisions even now, those decisions can now be overturned by the board of directors if necessary.
"While Leadership is the decision-maker, the Board of Directors holds the right to reverse decisions," OpenAI said, detailing its new framework model.
Microsoft Corp.-backed (NASDAQ:MSFT) OpenAI's primary focus with this new framework is on a critical topic – AI safety. Instead of assigning this important responsibility to just one person, the company has split it into three groups – Safety Systems, Preparedness, and Superalignment.
Safety Systems will look after current models, while Preparedness will work on frontier or next-generation models.
The Superalignment team, on the other hand, is tasked with the most challenging and futuristic job – preparing for artificial general intelligence (AGI).
In the end, OpenAI's leadership team, which includes Altman, will be tasked with taking decisions. However, if this team steps out of its bounds or goes against the overall mandate set by OpenAI, the board has now been empowered to veto those decisions.
Essentially, the board now has a failsafe to undo any decision that endangers the safety of the company or its stakeholders.
OpenAI notes that this framework is a work in progress and it comes with a "beta" tag as well, which means it will be continually worked upon and tweaked as need be.
Why It Matters: Empowering the board in a publicized safety framework is a move to ensure that Altman is still held accountable.
This is not controversial nor is it a power move against the co-founder, given that Altman has said it publicly earlier.
"No one person should be trusted here. I don't have super-voting shares. I don't want them. The board can fire me. I think that's important," Altman said earlier.
However, the timing is what makes this framework's release interesting. It comes a month after the company's board fired Altman unceremoniously. He was reinstated a few days later following a massive outcry from almost all of OpenAI's employees.