Italians may soon lose access to the popular artificial intelligence (AI) language model, ChatGPT, as Italy’s Privacy Guarantor orders the service to be blocked over concerns that OpenAI, the company behind ChatGPT, is violating the European Union’s General Data Protection Regulation (GDPR) with its data handling practices.
The regulator claims that OpenAI’s bulk collection of data for training ChatGPT’s model lacks a “legal basis,” while the sometimes-inaccurate results suggest that the generative AI is not processing data correctly. Furthermore, officials are particularly concerned about a flaw that leaked sensitive user data last week.
OpenAI’s alleged failure to protect children is also a significant concern for the regulator. Although the company states that ChatGPT is meant for people over the age of 13, officials claim that there are no age checks to prevent children from seeing “absolutely unsuitable” answers.
The Privacy Guarantor is giving OpenAI 20 days to address these issues. Failure to comply could result in a fine of up to €20 million (about $21.8 million US) or a maximum four percent of its annual worldwide turnover.
OpenAI’s ChatGPT privacy policy states that trainers can use conversation data to improve the AI, but that it also aggregates or anonymizes that data. The company’s terms forbid use by children under 13, while the policy claims that the company does not “knowingly” gather personal info from underage users.
We have reached out to OpenAI for comment and will update this article with their response. In the meantime, Italians may need to find alternative language models as ChatGPT’s future in Italy remains uncertain.
