Italy Temporarily Blocks ChatGPT Over Privacy Concerns

    Italy Temporarily Blocks ChatGPT Over Privacy Concerns
    Last updated Apr 01, 2023
    Image credit: AFP/Getty Images [via TechCrunch]


    • On Friday, Italy's Data Protection Authority temporarily blocked the artificial intelligence (AI) software ChatGPT over data privacy concerns, making it the first Western country to take such action against the chatbot.
    • As justification for the move, the agency cited "the mass collection and storage of personal data," also noting that ChatGPT suffered a data breach on March 20 involving "users' conversations" and subscriber payment details.
    • The Italian regulator also expressed concerns over "a lack of age verification" system for underage children even though San Francisco-based OpenAI, the company behind ChatGPT, claims users under 13 shouldn't use the platform.
    • IDPA is giving OpenAI 20 days to communicate the measures it will take to comply with Italy’s data rules or face fines of up to €20M ($21.8M) or 4% of its annual global turnover.
    • It's unclear how Italy will ban the platform on a national level, as it's unlikely to affect applications from companies that already have licenses with OpenAI to use the same technology behind the chatbot.
    • Meanwhile, in an open letter recently issued by the Future of Life Institute, artificial intelligence experts, researchers, and backers — including Twitter CEO Elon Musk — called for an immediate pause on creating AI systems more powerful than GPT-4 for at least six months, citing "profound risks to society and humanity."


    Narrative A

    Though OpenAI says isolated incidents are to blame for data breaches, the company may have had a security problem before the March 20 hack. ChatGPT's data scraping methods and its potential vulnerability to malware are among the justified and serious concerns being raised by public and private sector entities alike.

    Narrative B

    Despite what's being reported, to blame carelessness for the breach is flat-out wrong. An extremely small number of users were briefly exposed to the hack, but OpenAI quickly patched up the bug. ChatGPT is no more a threat to users' data than any other platform, and the whole story must be told before fear tactics replace common sense.

    Nerd narrative

    There's a 13% chance that OpenAI will limit interactions with GPT-4 on ChatGPT due to model misbehavior before the end of May 2023, according to the Metaculus prediction community.

    Sign up to our newsletter!