Ahead of OpenAI CEO Sam Altman’s four-day banishment, two people familiar with the case stated that several staff researchers submitted a letter to the board of directors warning of a significant artificial intelligence finding that they said could threaten humanity.
According to the two sources, the previously undisclosed letter and AI algorithm were critical developments prior to the board’s dismissal of Altman, the poster child of generative AI. Prior to his triumphal return late Tuesday, more than 700 staff vowed to quit in solidarity with their ousted leader, joining supporter Microsoft (MSFT.O).
The letter was highlighted as one element among a wider list of grievances by the board that led to Altman’s firing, including concerns about commercializing advancements before fully appreciating the ramifications. Reuters was unable to obtain a copy of the letter for review. The letter’s authors did not respond to calls for comment.
After being approached by Reuters, OpenAI, which declined to comment, confirmed a project dubbed Q* and a letter to the board before the weekend’s events, according to one of the sources. According to an OpenAI spokesman, the mail was sent by long-time CEO Mira Murati and alerted colleagues to particular media reports without commenting on their veracity.
OpenAI’s potential breakthrough in the quest for artificial general intelligence
According to Reuters, some at OpenAI believe Q* (pronounced Q-Star) could represent a breakthrough in the startup’s hunt for artificial general intelligence (AGI). AGI is defined by OpenAI as autonomous systems that outperform humans in the most economically valuable tasks.
The new model was able to solve some mathematical problems given massive computing capabilities, the person said on condition of anonymity because the individual was not authorized to talk on behalf of the company. Despite just performing math at the grade-school level, acing such exams made researchers highly confident about Q*’s potential success, according to the source.
Reuters was unable to independently verify the researchers’ claims about Q*’s capabilities.
‘The veil of ignorance’
Math, according to researchers, is a frontier of generative AI development. At the moment, generative AI is good at writing and language translation by statistically anticipating the next word, but answers to the same inquiry can differ greatly. However, mastering the capacity to solve math — where there is only one correct answer means that AI will have stronger thinking abilities similar to human intelligence. AI researchers believe that this might be applied to novel scientific studies.
In contrast to a calculator, which can only do a limited number of calculations, AGI can generalize, learn, and comprehend.
Researchers highlighted AI’s prowess and possible danger in their letter to the board, according to insiders, without elaborating on the specific safety issues raised in the letter.
Computer scientists have long debated the dangers posed by extremely intelligent robots, such as whether they would determine that destroying humanity is in their best interests.
Researchers have also identified work by an “AI scientist” team, the existence of which has been confirmed by various sources. One of the sources claimed the group, which was formed by integrating earlier “Code Gen” and “Math Gen” teams, was investigating ways to optimize existing AI models to increase their thinking and eventually perform scientific work.
OpenAI board ousts CEO Sam Altman amidst hints of major advances
Altman led efforts to make ChatGPT one of the fastest-growing software programs in history, attracting investment and computer power from Microsoft in order to come closer to AGI. This month, in addition to announcing a plethora of new tools in a demonstration, Altman last week teased at a summit of world leaders in San Francisco that he believed major advances were in sight.
“Four times now in the history of OpenAI, the most recent time was just in the last couple weeks. I’ve gotten to be in the room when we sort of push the veil of ignorance back and the frontier of discovery forward, and getting to do that is the professional honor of a lifetime,” he said at the Asia-Pacific Economic Cooperation Summit. A day later, the board fired CEO Altman.