OpenAI is still not training GPT-5, months after the Microsoft-backed startup pledged to not work on the successor to GPT-4 “for some time” after many industry executives and academics expressed concerns about the fast-rate of advancements by Sam Altman’s large language models.
“We have a lot of work to do before we start that model,” Altman, the chief executive of OpenAI, said at a conference hosted by Indian newspaper Economic Times. “We’re working on the new ideas that we think we need for it, but we are certainly not close to it to start.”
In late March, more than 1,100 signatories, including Elon Musk and Steve Wozniak, signed an open letter that calls on “all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4.”
Weeks later, Altman said that the letter was “missing most technical nuance about where we need the pause,” but asserted that OpenAI had not started training GPT-5 — and didn’t plan to do so for “some time.”
Altman on Wednesday pushed back again on the concerns from some of the most vocal voices on AI, saying the startup was already evaluating potential dangers with more meaningful measures such as external audits and red-teaming and safety tests.
“When we finished GPT-4, it took us more than six months until we were ready to release it.”
Earlier in the interview, Altman also said that OpenAI was against regulating smaller AI startups. “The only regulation we have called for is on ourselves and people bigger,” Altman said.
Altman’s trip to India is part of his attempt to aggressively meet with lawmakers and industry players globally and build confidence in OpenAI’s willingness to work with regulators. In his meetings, Altman is proactively urging lawmakers to put serious thinking into the potential abuse and other downside of AI proliferation so that guardrails could be put in place to minimize any unintended accidents.
Comment