ChatGPT faces ban in Italy over privacy concerns

Artificial intelligence tool ChatGPT was temporarily banned in Italy on Friday, the first known instance of the chatbot to be blocked by a Western government due to privacy concerns.

Italian data protection authority said OpenAI, the California company that makes ChatGPT, illegally collected personal information from users and lacked an age verification system to prevent minors from being exposed to unauthorized material.

The order is a sign of the policy challenges that arise for the developers of advanced AI after the release of ChatGPT. The program has stunned users with its ability to compose essays, engage in human conversations, and perform more complex tasks such as writing computer code, but it has raised alarms about the spread of misinformation, the impact on employment, and broader risks to society.

This week, more than 1,000 technology leaders and researchers called for a moratorium about the development of the most advanced AI systems, so that security policies can be implemented. The Center for AI and Digital Policy, an advocacy group that pushes for ethical use of technology, has early the US Federal Trade Commission to prevent OpenAI from releasing new commercial versions of ChatGPT.

In Italy, regulators have told OpenAI to block Internet users in the country from accessing ChatGPT until the company provides additional information. The company has 20 days to provide the agency with the material and possible solutions before a final decision can be made on the future of the product in the country.

Regulators cited a data breach on March 20 that exposed conversations and some users’ payment details. The agency said OpenAI could be fined up to 20 million euros (about $22 million) or 4 percent of its global annual revenue.

OpenAI declined to comment.

As of 5pm Friday in Italy, word had not yet reached the chatbot that it would be blocked in the country. When asked by a user there if it would be banned in Italy over privacy concerns, ChatGPT replied “there shouldn’t be any concerns.”

“I am a language model of artificial intelligence that can be accessed from anywhere in the world as long as there is an internet connection,” said the chatbot.

Emma Bubola reporting contributed.