Italy temporarily bans ChatGPT over privacy concerns

Italy has become the first Western country to ban the popular AI chatbot, ChatGPT due to concerns over privacy. 

Used by millions since its launch in November 2022, the AI answers user questions using natural, everyday language. Created by US Start-up OpenAI, ChatGPT has seen significant investment from Microsoft, which said it would embed the technology in its Office apps, such as Word and Excel.

US Start-up OpenAI created ChatGPT, which can answer questions using natural, everyday language. Microsoft has invested significantly in ChatGPT and has said it will embed the technology in its Office apps, such as Word and Excel.

Despite Open AI saying it complies with privacy laws, Garante, Italy’s Supervisory Authority, has blocked OpenAI’s chatbot and is investigating whether it complies with GDPR requirements. This block comes after a recent cyber breach which led to users being shown extracts of other users’ conversations as well as information such as their first and last names, billing addresses, credit card types, credit card expiration dates and the last four digits of their credit cards.

Storage of personal data

There is a concern around the mass collection and storage of personal data to train the algorithms, and Garante says the system has no legal basis for collecting this data. Garante also says that Open AI fails to verify the age of users, potentially exposing minors to ‘unsuitable answers’ from the chatbot given their relative lack of development and understanding. OpenAI risks facing a fine of €20 million, or 4% of its global annual revenue, if it doesn’t address how it will address the concerns within 20 days.  

The Irish Data Protection Commission is following up with the Italian authorities to understand the ban. The ICO said it is ready to ‘challenge non-compliance with data protection laws.

Regulation is necessary

Consumer advocacy group, BEUC, called on EU and national authorities to investigate ChatGPT and similar chatbots as there is not enough protection from the damage AI can cause. The EU is currently working on legislation relating to AI, but BEUC is concerned this could take years to take effect, leaving consumers at risk. ChatGPT is already blocked in a number of countries, including China, Iran, North Korea and Russia.

OpenAI confirmed it had disabled ChatGPT for Italian users but is looking forward to making it available again in Italy ‘soon’. OpenAI said it also believes AI regulation is necessary and will work closely with the Garante to provide further information on how the systems are built and used. OpenAI is working to reduce personal data in training AI systems as such systems should ‘learn about the world, not private individuals.’  

There have been calls for AI to face regulation, but it’s proving hard for governments and regulatory bodies to keep up with the speed at which technology and AI progress. It’s likely we’ll see steps to advance the regulation following pressure from groups such as BEUC.  

If you’d like to speak  about any issues raised in this article or would like help reviewing your own data protection policies, please contact Alex Craig on 0191 2117911 or email [email protected]