OpenAI proposed license to regulate AI, Baidu urges self content moderation

Jingyue Hsiao, DIGITIMES Asia, Taipei 0

Credit: AFP

As AI capabilities continue to advance, potentially disrupting most aspects of human life, OpenAI CEO Sam Altman proposed a licensing system requiring AI platforms to comply.

Multiple media, including the New York Times, Washington Post, and Bloomberg, reported that the United States Senate held a first major hearing on AI, with Sam Altman, OpenAI CEO, Christina Montgomery, IBM's chief privacy and trust officer, and Gary Marcus, a scholar, testifying before Congress.

According to the reports, Altman suggested forming a new agency that licenses any effort above a certain scale of capabilities to ensure compliance with safety standards, citing threats such as disinformation and lost jobs.

Bloomberg quoted Altman saying that while AI has the potential to solve humanity's biggest problems, the technology can cause significant harm to the world, adding that if this technology goes wrong, it can go quite wrong.

On the other hand, Montogomery said that efforts also focus on transparency so that people know when they are not interacting with a human but with an AI system. Participants in the hearing also talked about issues such as artist rights when AI is revolutionizing the music industry by generating music after training.

It is unclear how the potential regulations will impact the AI industry. New York Times said the US Congress has been discouraging, with dozens of privacy, speech, and safety bills failing to receive bi-partisan consensus. Still, Altman believes the US has enough potential to influence the global consensus on AI regulation as the US designs and controls much of the AI supply chain.

Meanwhile, instead of concerning the lost jobs and artist rights, China-based Baidu, which launched Ernie Bot in March, focused more on socially and politically sensitive disinformation. Nikkei Asia reported that Robin Li, Baidu CEO and founder, said in a call with analysts that they have to make sure AI will not hallucinate. He further said that Baidu continues to talk with relevant authorities and keep them informed about the latest technology trends, and organizations without significant experience in delivering suitable online content or a history of working closely with regulators will encounter substantial obstacles when attempting to provide scalable generative AI solutions in China.