Italy Bans ChatGPT Over Personal Data Risk
Italian data regulator bans ChatGPT in country and opens probe into whether chatbot’s data collection complies with GDPR
Italy’s data regulator has blocked ChatGPT in the country and launched an investigation into the Microsoft-backed chatbot’s use of personal data.
The Garante regulator said there was concern about the massive amounts of data collected by ChatGPT from its users.
It said there was no legal basis to justify “the mass collection and storage of personal data for the purpose of ‘training’ the algorithms underlying the operation of the platform”.
Italy is the first Western country to ban OpenAI’s ChatGPT, which is blocked in countries including China, North Korea, Iran and Russia.
Data breach
The regulator noted a 20 March software bug in the chatbot that exposed portions of users’ conversations and payment information to other users for about nine hours.
The breach confirmed previous warnings from industry watchers that sensitive information entered into the chatbot might be at risk.
The Garante added that since ChatGPT has no age-verification mechanism it “exposes minors to absolutely unsuitable answers compared to their degree of development and awareness”.
It said OpenAI had 20 days to respond to its concerns or it would face a fine of 20 million euros (£18m) or up to 4 percent of its annual revenues.
AI regulation
OpenAI said it had blocked ChatGPT in Italy, adding that it believes it complies with GDPR and other data protection laws.
It said it worked to reduce the personal data it uses in training AI systems.
“We also believe that AI regulation is necessary — so we look forward to working closely with the Garante and educating them on how our systems are built and used,” the company said.
Italy in February banned Replika.ai, which is powered by the same system behind ChatGPT.
Compliance risk
Security firm Cyberhaven in February estimated that sensitive data makes up 11 percent of what company employees enter into ChatGPT, creating compliance risks for firms that use it.
UK data breach law firm Hayes Connor said because Large Language Models (LLMs) of the kind that power ChatGPT are in their “infancy stages” companies using them are “in unchartered territory in terms of GDPR compliance”.
“Businesses that use ChatGPT without proper training and caution may unknowingly expose themselves to GDPR data breaches, resulting in significant fines, reputational damage, and legal action taken against them,” said Hayes Connor legal director Richard Forrest.
“As such, usage as a workplace tool without sufficient training and regulatory measures is ill-advised.”
