Italy has become the first country in the European Union to pass a national law regulating artificial intelligence, following the implementation of the bloc-wide AI Act. Now, children under 14 years old need parental consent to access AI systems.
Protecting children from AI is currently top of mind among politicians after a string of recent incidents involving chatbots. OpenAI is being sued by the parents of a 16-year-old boy who died by suicide. They claim ChatGPT reinforced his darker thoughts and offered information about how he could end his life.
In April, it was found that Meta’s AI chatbots, accessible through Facebook, Instagram, and WhatsApp, could engage in explicit conversations with minors if framed as role-playing. Plus, up until last month, Facebook’s parent company was actively allowing its chatbots to engage in “romantic or sexual” chats with children.
AI misuse in Italy could mean jail time, while some AI art can be protected by copyright law
The new law, approved by parliament on Wednesday, introduces prison sentences of between one and five years for AI misuse, such as spreading deepfakes, with harsher penalties for crimes like identity theft and fraud committed with AI.
The law defines a number of sector-specific safeguards, such as requiring doctors to make the final decisions even when using AI to aid in diagnosis and treatment, prohibiting judges from outsourcing their decisions to AI, and mandating that employers inform workers when AI tools are being used.
When it comes to the controversial topic of copyright, human-authored works created with the help of AI can be protected by copyright law if they show intellectual effort. At the same time, AI-driven text and data mining are only allowed for non-copyrighted content or scientific research.
Nations and tech companies worldwide have continuously debated how to balance artists’ rights, such as control over the use of their work and fair compensation, with the societal benefits of AI trained on vast amounts of data. The lack of clarity around permissions has already led to numerous lawsuits against the likes of OpenAI, Meta, and Anthropic.
Italy has provided some answers for artists and AI developers, but since many models operate internationally, the legislation could just as easily create new layers of confusion.
The country’s new law authorises up to €1 billion ($1.18 billion) available through a state-backed venture capital fund to support companies in AI, cyber security, quantum, and telecoms. Its enforcement will be led by the Agency for Digital Italy and the National Cybersecurity Agency.
The country has had AI in its crosshairs for a while
Italy has something of a reputation for scrutinising AI.
In January 2024, the country’s data protection authority accused OpenAI’s ChatGPT of violating GDPR by processing personal data without a proper legal basis, leading to a temporary suspension. With its new law, it has shown it is not waiting for the complete enactment of the EU AI Act in 2030 for AI regulation within its borders.
Coincidentally or not, Italy’s AI law was authorised just a day after Mario Draghi, former European Central Bank President and author of a 2024 report criticising the EU’s weak AI competitiveness, suggested that the EU’s AI Act is put on pause “until we better understand the drawbacks.”
The Federal Trade Commission has launched a sweeping inquiry into AI chatbots, pressing seven tech giants over potential risks to children and teens.