OpenAI Inc. Chief Executive Officer Sam Altman plans to tell Congress that artificial intelligence technology is becoming so powerful that the US and other governments should consider imposing regulation on products "above a crucial threshold of capabilities.”
Altman, who will testify for the first time since OpenAI’s ChatGPT exploded in popularity, plans to say that artificial intelligence companies should be subjected to a "combination of licensing or registration requirements” for the release of advanced models, according to his prepared remarks, which don’t specify what that threshold would look like.
It’s vital "that AI companies - especially those working on the most powerful models-adhere to an appropriate set of safety requirements,” he plans to say. A copy of his prepared remarks to a Senate panel was obtained by Bloomberg News.
Congress is debating the potential and pitfalls of artificial intelligence as products like ChatGPT raise questions about the future of creative industries and the ability to tell fact from fiction.
Even as Tuesday’s hearing will discuss safeguards to prevent this technology from being abused, Senator Richard Blumenthal, a Connecticut Democrat and chair of the Judiciary Subcommittee on Privacy, Technology and the Law, said he’d also ask what US companies need to ensure supremacy in the field.
"We need to be the world leader in AI,” Blumenthal said in a telephone interview Monday. "We can’t allow China to surpass us.”
Altman will tell the senators that he’s engaged with policymakers worldwide to help them understand artificial intelligence and what guardrails should be in place. He will also urge the imposition of rules that would apply no matter where the products are deployed.
"We are not alone in developing this technology,” Altman plans to say. "It will be important for policymakers to consider how to implement licensing regulations on a global scale and ensure international cooperation on AI safety.”
San Francisco-based startup OpenAI has captured the world’s attention with products built on its large language models - AI systems that ingest enormous volumes of digital text from the internet, and use that material to train software that predicts and generates content on its own when given a prompt or query.
Millions of people have experimented with ChatGPT, and other companies and have released similarly capable chatbots. Microsoft, a major OpenAI investor, rolled out its Bing chat powered by OpenAI’s technology, while Alphabet Inc.’s Google created one called Bard.
Christina Montgomery, chief privacy and trust officer for IBM Corp., also plans to testify Tuesday. According to her prepared remarks, she plans say that regulation should focus on how the technology is used, leaving room for innovation in constructive applications like modernizing electricity grids and preventing fraud.
Regulators should stick to this approach, despite the "explosion of generative AI systems” in recent months, she plans to say.
The European Union parliament is wrestling with the same question as it finalizes its AI Act. Policymakers in Brussels are now considering whether to expand the scope of that regulation to include generative AI products like ChatGPT. – Bloomberg