There's been a whirlwind of artificial intelligence news over the past week, but OpenAI and Sam Altman weren't the only ones generating headlines.
AI researcher Helen Toner spoke candidly about what she observed during her time as a board member at OpenAI, including what led to the committee's decision to fire Sam Altman back in November.
Toner made the comments during an appearance on the TED AI Show, a podcast produced by the company known for TED Talks. You can find the full episode here.
"For years, Sam had made it really difficult for the board to actually do that job by withholding information, misrepresenting things that were happening at the company, in some cases outright lying to the board,” Toner said.
Among other things, the board wasn't given any notice that ChatGPT would be released in November 2022, Toner said, and Altman also never disclosed to the board that he owned the OpenAI Startup Fund.
"We learned about ChatGPT on Twitter," Toner said.
Rules and regulations
In the legislative realm, California Governor Gavin Newsom indicated that he is trying to find the right balance between regulations and fostering innovation.
“When you have the inventors of this technology, the godmothers and fathers, saying: ‘Help, you need to regulate us,’ that’s a different environment,” Newsom said during an AI event in San Francisco on Wednesday, Politico reported.
There are several AI-related bills currently being discussed by state lawmakers, but Newsom hasn't publicly commented on whether he intends to signs any of them.
Data deals
On Wednesday, OpenAI announced agreements with two more news media outlets: Vox Media and the Atlantic. Both are multi-year deals, though the exact terms weren't disclosed, and OpenAI will be allowed to license content from Vox and the Atlantic to train its large language models, Axios reported.
OpenAI will also be allowed to surface content from Vox and the Atlantic with citations in response to user inquiries, and both publishers will have access to OpenAI's technology.
Trust and safety
OpenAI says it formed a new "safety and security committee" that will ostensibly replace the "superalignment" team which dissolved amid the resignations of two safety researchers: Ilya Sutskever and Jan Leike.
The superalignment team was tasked with researching ways to ensure humans would be able to "steer and control" an advanced general intelligence system, or AGI, that was smarter than us. Sutskever and Leike led that team.
Leike was quickly scooped up by rival Anthropic, but Sutskever hasn't announced his next move quite yet.
The new committee is made up of nine OpenAI insiders, including Altman and three other board members, which could raise some eyebrows.
And just a couple of weeks after unveiling an update to GPT-4, OpenAI said it was training a new foundational model to replace the current model, though it isn't likely to be released till sometime next year at the earliest, the New York Times reported on Tuesday.
The update to GPT-4, though, triggered outrage from actress Scarlett Johansson who alleged that OpenAI may have mimicked her voice for one of the new voice chat profiles unveiled in the update, dubbed GPT-4o, or omni. OpenAI subsequently pulled the voice.
Funding rounds to know
Elon Musk raised $6 billion in a series B round for his OpenAI competitor, xAI. The round valued xAI at $24 billion and was backed by investors including Andreessen Horowitz and Sequoia Capital, Reuters reported. In the Bay Area, xAI has offices in San Francisco and Palo Alto.
