Europe, US urged to investigate type of AI powering systems like ChatGPT

European Union consumer protection groups urged regulators on Tuesday to investigate the type of artificial intelligence underpinning systems like ChatGPT, citing risks that leave people vulnerable and the delay before the bloc's groundbreaking AI regulations take effect.

In a coordinated effort, 13 watchdog groups wrote to their national consumer, data protection, competition and product safety authorities warning them about a range of concerns around generative artificial intelligence.

A transatlantic coalition of consumer groups also wrote to U.S. President Joe Biden asking him to take action to protect consumers from possible harms caused by generative AI.

Europe has led the world in efforts to regulate artificial intelligence, which gained urgency with the rise of a new breed of artificial intelligence that gives AI chatbots like ChatGPT the power to generate text, images, video and audio that resemble human work.

The EU is putting the finishing touches on the world's first set of comprehensive rules for the technology, but they are not expected to take effect for two years.

The groups called for European and U.S. leaders to use both existing laws and bring in new legislation to address the harms that generative AI can cause.

They cited a report by the Norwegian Consumer Council outlining dangers that AI chatbots pose, including providing incorrect medical information, manipulating people, making up news articles and illegally using vast amounts of personal data scraped off the internet.

The consumer groups, in countries including Italy, Spain, Sweden, the Netherlands, Greece and Denmark, warn that while the EU's AI Act addresses some of the concerns, they won't start applying for several years, "leaving consumers unprotected from a technology which is insufficiently regulated in the meantime, and developing at great pace."

Some authorities have already acted. Italy's privacy watchdog ordered ChatGPT maker OpenAI to temporarily stop processing user's personal information while it investigated a possible data breach. France, Spain and Canada also have been looking into OpenAI and ChatGPT.