OpenAI CEO, Sam Altman, arrives for a bipartisan Artificial Intelligence (AI) Insight Forum for all U.S. senators hosted by Senate Majority Leader Chuck Schumer (D-NY) at the U.S. Capitol in Washington, U.S., on September 13, 2023. REUTERS/Julia Nikhinson/File Photo
Nov 22 (Reuters) – Prior to OpenAI CEO Sam Altman’s temporary departure, a group of staff researchers sent a warning letter to the board of directors, detailing a groundbreaking artificial intelligence discovery that they claimed could pose a threat to humanity, according to two individuals familiar with the situation.
The unexpected letter, along with the AI algorithm named Q*, played a pivotal role in the board’s decision to remove Altman, the well-known figure behind generative AI technology, as outlined by the sources. Before Altman’s return on Tuesday, over 700 employees had threatened to quit and go join Microsoft, one of OpenAI’s backers, in solidarity with their ousted leader.
The sources pointed out that the letter was just one of several concerns raised by the board that led to Altman’s dismissal. Reuters was not able to obtain a copy of the letter, and there was no immediate response from the researchers who authored the letter when requested for comment.
OpenAI declined to provide a statement.
According to one source, Mira Murati, a long-time executive at the company, discussed the Q* project with employees and mentioned that a letter had been sent to the board prior to the recent events.
Following the publication of the story, a spokesperson for OpenAI stated that Murati had informed employees about the media reports but did not verify the accuracy of the information.
The creator of ChatGPT had been making strides with Q* (pronounced Q-Star), which several internal parties believe could signal a breakthrough in the startup’s quest for superintelligence, also known as artificial general intelligence (AGI), as revealed by one of the sources. OpenAI defines AGI as AI systems that surpass human capabilities.
With ample computing resources, the new model demonstrated the ability to solve certain mathematical problems, as stated by the individual who preferred to remain anonymous due to a lack of authorization to speak on the company’s behalf. While only displaying math proficiency at a grade-school level, this accomplishment generated a great deal of optimism among researchers regarding Q*’s potential future success, the source added.
Reuters was unable to independently confirm the capabilities of Q* as described by the researchers.
**This article includes reporting by:
Anna Tong and Jeffrey Dastin in San Francisco and Krystal Hu in New York; Editing by Kenneth Li and Lisa Shumaker**