Business Maverick
OpenAI leaders address staff concerns over equity, Scarlett Johansson
OpenAI CEO Sam Altman and several other executives answered questions about a nondisparagement contract for outgoing employees at a meeting on Wednesday, following a series of controversies involving the company in recent weeks.
The company has apologized for sections of a contract that tied outgoing employees’ equity to an agreement not to disparage OpenAI. On Thursday, it confirmed it had removed the wording from its standard exit contract and that it was releasing the majority of former employees from non-disparagement agreements.
In a post on X on Saturday, Altman apologised for the policy, calling it one of the few times he’s been “genuinely embarrassed” while running OpenAI.
At the Wednesday meeting, a regular weekly gathering with leadership and staff, Altman joined chief strategy officer Jason Kwon and chief operating officer Brad Lightcap in answering questions about the non-disparagement agreements, according to a person familiar with the matter who asked not to be identified discussing private information. Vox had written about the agreements days earlier.
OpenAI came to prominence in 2022 after the release of its ChatGPT chatbot. The startup is backed by Microsoft Corp., which invested $13-billion and has infused OpenAI’s technology into its software products in a competition for paying customers with Alphabet’s Google and others. OpenAI was valued at $86-billion in February in a deal to let its employees sell their stakes in the company.
During the meeting, OpenAI chief technology Officer Mira Murati also briefly discussed the situation surrounding allegations by Scarlett Johansson that the company used a voice “eerily similar” to her own in its AI voice assistant tool after she rejected an offer by Altman to be a part of the project.
Murati walked through the company’s timeline of hiring outside voice actors for the tool and how the voices were selected, the person said. Altman has said publicly that the voice was trained on an outside actor and was “never intended to resemble” Johansson. He said he has since apologised to the Hollywood star, saying that the company needed to “communicate better” on the topic.
After a period of relative calm since OpenAI’s boardroom drama in November, the company has been back in the news this month. Last week, chief scientist Ilya Sutskever, who played a key role when Altman was fired and reinstated, resigned along with researcher Jan Leike. Leike publicly shared his concerns about the company’s commitment to studying the long-term risks of AI. The safety team they worked on at the company no longer exists.
The whole AI concept and its operating platforms now appear to be slipping out of the control of human beings as it accelerates its own capabilities of building on itself. Scientists are now admitting to real concern as they are finding it increasingly difficult it tests to ascertain if they’re communicating with a human being or a machine. International cyber confrontation and warfare is already emerging and terrorist groups and countries like Russia and China are eyeball-to-eyeball with the US and the West in a Spy vs Spy race to develop and weaponise AI. Scary stuff.