Italy Restricts ChatGPT, Launches Investigation Over Privacy Concerns

MILAN/STOCKHOLM, March 31 (Reuters) – Italy’s data protection authority on Friday temporarily banned OpenAI’s ChatGPT chatbot and launched an investigation into alleged breaches of data collection rules for artificial intelligence applications.

The agency, also known as Garante, accused Microsoft Corp-backed (MSFT.O) ChatGPT of not checking the age of its users, who are supposed to be over 13 years old.

ChatGPT “has no legal basis to justify the large-scale collection and storage of personal data” to “train” the chatbot, Garante said. OpenAI said he could take corrective action within 20 days or be fined up to 4% of annual global turnover.

OpenAI did not immediately respond to requests for comment.

ChatGPT was still answering questions posted on the platform by Italian users on Friday night.

The company was informed of the decision on Friday morning, making it virtually impossible to suspend access in Italy on the same day, but a spokesperson for the agency expects to do so by Saturday. said.

“If they ignore the ban, authorities can impose fines,” the spokesman said.

Italy, which temporarily restricted ChatGPT’s use of domestic users’ personal data, became the first western country to take action against chatbots powered by artificial intelligence.

Chatbots are also not available in mainland China, Hong Kong, Iran, Russia, and parts of Africa where residents cannot create OpenAI accounts.

Since its release last year, ChatGPT has sparked a tech frenzy, prompting rivals to launch similar products and companies to integrate ChatGPT or similar technology into their apps and products.

The rapid development of this technology has caught the attention of several national legislators. Many experts say new regulations are needed to govern AI because of its potential impact on national security, jobs and education.

The European Commission, which is debating an EU AI law, may not intend to ban AI, tweeted European Commission Vice President Margrethe Vestager.

“No matter what #tech we use, we must continue to advance freedom and protect rights. I will,” she said. “Don’t throw away in a few years what took decades to build.”

The EC did not respond to requests for comment.

On Wednesday, Elon Musk and a group of artificial intelligence experts and industry executives said in an open letter, citing potential risks to society, that OpenAI would not develop a system more powerful than its newly launched GPT-4. Asked to stop for a month.

OpenAI does not provide details on how the AI ​​model is trained.

Johanna Björklund, an AI researcher and associate professor at Umeå University in Sweden, said: “If you do AI research, you have to be very transparent about how you do it.”

According to a UBS study published last month, ChatGPT is estimated to have reached 100 million monthly active users in January, just two months after its launch, making it the fastest-growing consumer in history. It is an application for people.

Reporting by Elvira Pollina from Milan, Supantha Mukherjee from Stockholm, Crispian Balmer from Rome and Foo Yun Chee from Brussels. Written by Alvise Armellini and Alessia Pé, edited by Gavin Jones and Richard Chang

Our standards: Thomson Reuters Trust Principles.