OpenAI CEO Sam Altman has confirmed that the Trump administration reviewed the company's upcoming artificial intelligence model Astra before its expected release, describing the voluntary process as productive.
Axios reported that Altman said OpenAI participated in the review despite the process not being mandatory. “But we of course did it,” he said, referring to the administration's review of Astra.
Astra is significant for OpenAI because it is the company's first model to be designated as having a critical level of cyber capability. That classification means the model is subject to stronger safeguards during development and before it is released.
The US administration has established a voluntary framework under which government officials can review some of the most advanced AI models ahead of their public launch. The process can potentially give officials access to a model for as long as 30 days.
However, the White House has not made the framework public, leaving details of how the reviews are conducted largely unclear.
ALSO READ | OpenAI Discovers More AI Agent Containment Breaches During Hacking Probe: Report
Altman said closer cooperation between AI companies and governments will become increasingly important as the technology advances.
“As these models get to a quite significant level of capability,” he said, engagement with safety institutes in the US, UK and other countries will become more important.
Altman described Astra as a significant step forward and said OpenAI expects to continue developing increasingly capable models.
His comments came after he attended the G20 Innovation Ministerial, where he spoke alongside US Commerce Secretary Howard Lutnick. Altman described OpenAI as the pragmatic centrists and said people should remain at the centre of economic and social decision-making.
The discussion also touched on the risks surrounding AI. Nvidia CEO Jensen Huang cautioned against fear-driven messaging, warning that countries could fall behind if they fail to take advantage of the technology.
Altman, however, argued for a balanced approach. He said excessive doomerism around AI risks could be harmful, but added that blind optimism without acknowledging those risks was equally problematic.
ALSO READ | OpenAI Says It Could Have Reacted Sooner To Prevent AI Hack Of Hugging Face
OpenAI and Anthropic have also faced criticism that their backing of certain AI regulations could strengthen the position of leading AI companies and make it harder for smaller rivals to compete.
Altman said he had expected Congress to establish a basic regulatory framework for advanced AI after he testified before lawmakers in 2023. That has yet to happen, he said, partly because the technology is evolving too quickly.
For OpenAI, the next phase is likely to involve the release of Astra and other increasingly powerful models as governments around the world continue to grapple with AI safety and regulation.
Altman said the company has much, much, much more capable models coming soon, signalling the rapid pace of AI development.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.
