Advertisement

ChatGPT banned in Italy over privacy and age verification concerns

ChatGPT is an AI or artificial intelligence programme  (Alamy / PA)
ChatGPT is an AI or artificial intelligence programme (Alamy / PA)

ChatGPT is now banned in Italy.

The country’s data protection watchdog said on Friday (March 31) that it has ordered use of the bot to be halted immediately over what they say is its “unlawful collection of personal data”. It also claimed the AI tool lacked systems for verifying the age of minors.

Alongside the immediate ban, the watchdog has opened an investigation into OpenAI, the US-based research lab that created ChatGPT.

The clampdown comes after the viral chatbot suffered a data breach on March 20, which saw the conversation histories and payment information leaked for users of its premium subscription service. At the time, OpenAI CEO Sam Altman said he regretted the leak, and the company fixed the problem.

Italy is the first western country to outright ban the bot. The action is part of a broader move to limit the spread of public AI systems, the prevalence of which have grown since the launch of ChatGPT last December. OpenAI licenses the underlying tech for its bot to other companies, including Microsoft, Snapchat, and Slack.

Meanwhile, the UK issued a white paper on AI on Wednesday (March 29). With the aim of striking a balance between regulation and innovation, the Government plans to use existing regulators in different sectors rather than giving responsibility for AI governance to a new single regulator.

More than 1,100 tech experts, including Elon Musk, also recently signed an open letter demanding a pause on powerful AI systems in the absence of regulatory oversight.

Italy’s data watchdog said OpenAI does not have a legal basis to justify its mass collection and storage of personal data. The company uses the information to train its AI systems, which include ChatGPT and AI-powered image generator Dall-E.

The watchdog added that although OpenAI says that ChatGPT is only accessible to those aged over 13, it does not have any filter to check the age of users. This, it added, “exposes minors to absolutely unsuitable answers compared to… their degree of development and self-awareness.”

OpenAI now faces a fine of up to €20 million (£17.6), or up to four per cent of its annual global turnover, if it does not implement the changes requested by the Italian watchdog within 20 days. In February, the regulator also banned a chatbot known as Replika over similar grounds.