During his visit to India, OpenAI CEO Sam Altman said he would not pause the development of newer iterations of ChatGPT. This comes amid calls by tech luminaries to hit the breaks on AI research and development and allow governments and institutions to build better regulatory frameworks to prevent misuse of the technology.
“A pause sounds good but I don’t think it works in practice,” Altman said on Thursday while speaking at a conference in IIIT-Delhi. “Norms like external audits, red teaming and testing are much better ways [to prevent misuse].”
Red teaming involves testing an organisation’s security posture by simulating real-world cyber-attacks.
Altman revealed that he will make investments in AI in India and has been talking to a few companies about the same. However, while the exact details are not clear, he said the investments will come from OpenAI Startup Fund I, which closed a $175 million fundraise last month.
Attending multiple events in the national capital, Altman had to repeatedly assure that OpenAI’s GPT models are harmless and are incapable of “sci-fi like capture of humanity,” and there is little to no risk.
He also highlighted the need for regulations to ensure bigger companies don’t get too powerful and conduct AI research in a transparent manner. Altman also reiterated that he is against regulating smaller companies working on language models as it would hurt the progress of AI.
“The world can come together for important things…This is an existential risk. If the governments don’t get their act together, we will try our hardest to get every company to cooperate, but we can’t get every company to cooperate. We would at least like to ask for a dream world and, if we can’t have that, we will get the companies that wanna play ball together and do our best,” Altman added.
The OpenAI CEO, who was earlier a partner at Y Combinator and founded social network Loopt when he was 19, also expressed regret advising startup founders on approaching AI as he broke a lot of his own rules when he was building OpenAI.
“But some principles haven’t changed. As we said back in Y Combinator—build something people need—and we would like to stick with thought at OpenAI,” he said
Speaking at a presser, Sam urged the Indian government to make AI the centre of conversation at the G20 Leaders’ Summit in September this year. The founder also met G20 sherpa Amitabh Kant at his office where they discussed “how emerging economies can leverage generative AI to improve quality of life” among other things.
On his visit to India, NASSCOM President Debjani Ghosh said in a LinkedIn post, “I hope that when Sam comes back—he does keep some time for real discussions with the tech community as his experience and insights can help us a lot.”.
At an ET event, Rajan Anandan, Managing Director of Peak VX (formerly Sequoia India) asked Altman for suggestions to startups building foundational models for AI. The OpenAI CEO simply said, “There is no point in competing with us. We are way ahead of everyone,” and challenged founders to prove him wrong.
He also met Prime Minister Narendra Modi and reportedly discussed India’s early adoption of ChatGPT.
