Jan Leike believes that OpenAI “is taking on an enormous responsibility on behalf of all humanity.” And Leike’s opinion is important, since he was, until now, in charge of analyzing that responsibility and neutralizing its risks. In a long thread on X, the security manager of the company that created ChatGPT has said that it prioritizes the creation of new and attractive products over security. And for all these reasons he has justified his departure.
Leike was co-director of the super-alignment department, along with Ilya Sutskever, who left the company a few days ago. His subordinates, William Saunders Leopold Aschenbrenner, complete the list of casualties. OpenAI thus dismantles its security team, which was responsible for ensuring that powerful artificial intelligence systems respected human values and objectives.
Yesterday was my last day as head of alignment, superalignment lead, and executive @OpenAI.
— Jan Leike (@janleike) May 17, 2024
The entire tech world was looking at the San Francisco company, but for other reasons. This week it launched ChatGPT-4o, the free and improved version of its conversational chatbot. During the presentation, many were struck by the attendee’s ability to flirt with the event presenters and the references made by the company’s CEO, Sam Altman, to the film Her, in which a man falls in love with a conversational chat. dystopia as marketing.
In this context, a cascade of casualties has occurred. That of Ilya Sutskever has been the most popular, since he was one of the founders and, until recently, a member of the hard core of the technology company. Sutskever wrote in X’s post announcing his departure that he was confident OpenAI will “build a safe and beneficial AGI” under his current leadership.
Sutskever represented the most conservative wing of the company and for that reason, last November, he was one of those who attempted the sudden dismissal of Altman, whom they accused of excessive recklessness. Altman regained his position, in a business drama that continues to have consequences to this day. This idea is gaining strength now with the cascade of resignations: in the company there would be a series of engineers who advocated waiting before marketing such advanced models. But they are abandoning it.
OpenAI is expected to soon launch ChatGPT Search, the AI-supported search engine that would turn the search engine sector upside down, until now controlled by Google, and the long-awaited ChatGPT5, the new version of its star tool. The company has also developed other products such as the Dall-E image generator and the Sora video generator.
With all these new developments on the market and an open debate about the risks of artificial intelligence, Leike’s reflections are clear and do not leave his former company in a good place. “I have been at odds with OpenAI management over the company’s core priorities for quite some time,” he explained. “Until we finally reached a breaking point.” The engineer believes that OpenAI should invest more resources in issues such as security, social impact, confidentiality and security for its next generation of models. “These problems are very difficult to solve, and I worry that we are not on the right path,” he wrote.
On the same social network, Altman has responded to his former employee, expressing gratitude for the service provided over the years and claiming to be “very sad” about his departure. In a message that he has promised to expand on in the coming days, he has acknowledged that the engineer’s criticisms are justified: “He is right, we have a lot left to do and we are committed to doing it.”
You can follow EL PAÍS Technology in Facebook and x or sign up here to receive our weekly newsletter.
#OpenAI #puts #flashy #products #security #worker