05/17/2024 – 23:55
OpenAi confirmed this Friday (17) that it has dissolved a team dedicated to reducing the long-term risks of artificial intelligence (AI), at a time when the company that developed ChatGPT faces a growing number of criticisms.
Weeks ago, the dissolution of the so-called “superalignment” group began, with the transfer of its members to other projects and research, reported the company, based in San Francisco.
Earlier, the director of this team responsible for security, Jan Leike, explained on the social network X that he resigned due to fundamental differences with senior management over the company’s priorities, in a dilemma between innovation and security.
“We have reached a breaking point,” said the engineer in charge of the group responsible for superalignment, that is, ensuring that a future “general AI”, one that can be as intelligent as man, is aligned with society’s values.
Sam Altman, the company’s co-founder and CEO, expressed sadness at Leike’s departure. “You’re right, we still have a lot to do (in alignment and safety research) and we are determined to do it,” he said.
The superalignment team was also led by Ilya Sutskever, co-founder of OpenAI, who announced his departure last Tuesday.
With ChatGPT, OpenAI launched the generative AI revolution, capable of producing content from a simple request in everyday language, which generated enthusiasm in Silicon Valley, but which, on the other hand, worries many analysts and regulators due to its security and data usage, from the United States to Europe.
This week, OpenAI presented a new version of this “chatbot”, which can now maintain fluid oral conversations with its users, a further step towards increasingly personal and powerful AI assistants.
#OpenAI #dissolves #future #risks #unit