Italy’s data protection watchdog, Guarantor for the Protection of Personal Data (aka Guarantor), has temporarily banned OpenAI’s ChatGPT service in the country, citing data protection concerns.
As such, the company has ordered it to stop processing your data immediately and intends to investigate whether it is unlawfully processing such data in violation of the EU’s General Data Protection Regulation (GDPR) law. says.
“No information is provided to users and data subjects whose data is collected by Open AI,” said Garante. “More importantly, there appears to be no legal basis for collecting and processing large amounts of personal data to ‘train’ the algorithms the platform relies on. ”
ChatGPT, which has an estimated 100 million monthly active users since its launch late last year, is wondering what or how it used to train its latest large-scale language model (LLM), GPT-4. You didn’t reveal how you trained.
That said, its predecessor, GPT-3, draws on texts sourced from books, Wikipedia, and Common Crawl, which maintains an “open repository of web crawl data that anyone can access and analyze.” increase.
Garante also noted that there is no age verification system to prevent minors from accessing the service, which could expose minors to “inappropriate” responses. Google’s own chatbot called Bard is only available to users over the age of 18.
Additionally, regulators questioned the accuracy of information surfaced by ChatGPT while highlighting a data breach the service suffered earlier this month, exposing some users’ chat titles and payment-related information. bottom.
Following this order, OpenAI will I was blocked Its generated AI chatbot prevents users with Italian IP addresses from accessing it. It also said it was issuing refunds to ChatGPT Plus subscribers in addition to pausing subscription renewals.
The San Francisco-based company further emphasized that ChatGPT is compliant with GDPR and other privacy laws. ChatGPT is already blocked in China, Iran, North Korea and Russia.
In a statement shared with Reuters, OpenAI said it was “actively working to reduce personal data when training AI systems like ChatGPT.”
OpenAI will notify Garante within 20 days of any action taken to achieve compliance. Failure to do so risks facing fines of up to €20 million or 4% of his global annual turnover, whichever is higher.
However, the ban is not expected to affect applications from other companies that employ OpenAI’s technology to power their services, such as Microsoft’s Bing search engine and Copilot products.
Become an Incident Response Pro!
Unlocking the Secrets of Bulletproof Incident Response – Master the 6-step process with Asaf Perlman, IR Lead at Cynet!
Don’t miss it – secure your seat!
The development has led Europol to believe that LLMs like ChatGPT are likely to generate malicious code and facilitate fraud, and “could provide new opportunities for criminals, especially for crimes involving social engineering.” It is also accompanied by the warning that the risk is high. style. “
This isn’t the first time AI-focused companies have been spotlighted. Last year, controversial facial recognition company Clearview AI was fined by multiple European regulators for scraping public photos of users without their consent to train identity matching services.
It also violates Australian, Canadian and US privacy laws, and several countries have ordered the company to delete all data obtained in such ways.
Clearview AI told BBC News last week that it had searched US law enforcement agencies nearly one million times, despite being permanently banned from selling a database of headshots in the US.
