Sam Altman (pictured), chief executive of OpenAI, the company behind ChatGPT, has told US legislators that artificial intelligence will cause “significant harm to the world” without urgent regulatory intervention.
However, Altman said that the firm would not rule out running ads on its service in the future despite his preference for a subscription-based business model.
“My worst fear is that we, the field, the technology, the industry, cause significant harm to the world,” Altman said.
“If this technology goes wrong it can go quite wrong, we want to work with the government to prevent that from happening.”
Of particular concern for Altman was generative AI’s ability to interfere with democratic processes, saying he was “nervous about it and a “significant area of concern” that needed rules and guidelines.
Senator Mazie Hirono said “In the election context, for example, I saw a picture of former President Trump being arrested by NYPD and that went viral,” she said, and pressed Altman on whether he would consider that faked image harmful.
Altman said that creators should make clear whether an image was generated rather than real and suggested that, in general, the US should consider licensing and testing requirements for the development of AI models.
Section 230, the clause in America’s Communications Decency Act 1996 that limits the liability of online publishers for content posted by users, was also up for debate at the hearing. Altman said that a new regime was needed in the era of AI.
“How we want to come up with a liability framework there is a super important question,” he said.
He also said that companies should have the right to say whether they wanted their data used for AI training, an idea that has garnered some backing from new media bigwigs in Australia and the US. Nine’s Mike Sneesby recently said that he wanted the News Media Bargaining Code to be expanded to cover AI companies.
On the question of advertising, Altman said he “wouldn’t say never” and that for some customers for whom there is no other way to access its services, it might be a plausible solution. However, he added that he “really like[s]” having a subscription-based model.
Of course, OpenAI is already in the advertising business to some extent. The company’s software has been integrated into Microsoft’s Bing search engine, which is already serving contextual and keyword-based adverts on the back of user inputs.
As it stands, ChatGPT is free to use though users may subscribe to the more powerful ChatGPT Plus system for US$20 per month (AU$30 per month). Google’s Bard AI system, meanwhile, is already serving ads.