OpenAI not training GPT-5

OpenAI, the Microsoft-backed startup, has announced that it is still not working on the development of GPT-5, the successor to GPT-4, despite earlier pledges. The decision comes after industry executives and academics expressed concerns about the rapid advancements made by Sam Altman’s large language models.

Speaking at a conference hosted by ET, OpenAI CEO Sam Altman stated, “We have a lot of work to do before we start that model. We’re working on the new ideas that we think we need for it, but we are certainly not close to it to start.” Altman emphasized that they are focused on further preparations before proceeding with GPT-5.

In late March, a letter signed by over 1,100 individuals, including Elon Musk and Steve Wozniak, called for a six-month pause on training AI systems more powerful than GPT-4. Altman acknowledged the concerns but clarified that the letter lacked technical nuance. He reaffirmed that OpenAI had not commenced training GPT-5 and had no plans to do so in the near future.

Altman addressed the concerns raised by AI experts, stating that OpenAI is actively evaluating potential risks through external audits, red-teaming, and safety tests. He highlighted the thoroughness of their process, explaining, “When we finished GPT-4, it took us more than six months until we were ready to release it.”

During the interview, Altman also expressed OpenAI’s stance against regulating smaller AI startups. He emphasized that the only regulation they advocate for is self-imposed and applicable to larger organizations. Altman’s visit to India is part of OpenAI’s proactive approach to engage with lawmakers and industry players worldwide, seeking to build trust and demonstrate their willingness to collaborate with regulators. He is urging policymakers to carefully consider the potential risks and drawbacks of AI proliferation, with the aim of implementing safeguards to minimize unintended accidents.