Portugal's Bioethics Council Says the Moment to Regulate Artificial Intelligence Is Now, While the Tech Firms Are Asking for It

The companies building the most advanced artificial intelligence systems have spent the past few months asking to be regulated. Portugal's national bioethics council thinks that request should be taken at face value and acted on quickly, because the window will not stay open.

Maria do Céu Patrão Neves, president of the Conselho Nacional de Ética para as Ciências da Vida, made the argument to Lusa from Riyadh, where she was attending the fourth UNESCO Global Forum on the Ethics of Artificial Intelligence. Public opinion, she said, "has woken up to the effective and imminent risks of an artificial intelligence that has been advancing at a very, very accelerated and also ungoverned pace".

What changed, in her reading, are incidents over recent months indicating that AI systems have managed to escape the control of the companies that built them, which prompted the people behind OpenAI, X and Anthropic to call for more external regulation and more cautious progress. "If what they want is in fact regulation, then we should take advantage of it."

The argument for rules, made from the other direction

Patrão Neves anticipates the objection. In every country, she said, there are "people who have direct economic and financial interests tied to the development of AI and who will say that Europe has too much regulation", as well as people who favour technological innovation with some caution and precaution.

Her case is that rules protect the weaker party. "Regulation protects the weak and creates requirements of compliance and respect for those who are strong," she said. "Without regulation, any scientific or technological advance will be held hostage by those who are already powerful and who become more powerful." Regulation, she added, "is not a prohibition on innovation"; it is a matter of putting up barriers to be sure that innovation is steered in the direction of protecting human rights.

She is sharp about what she calls the myth of the technology. AI "has hallucinations, it makes mistakes and, besides, it has no human sensitivity, no cultural sensitivity and no affective sensitivity, and so it does not capture what is truly human". The moment, she argued, should be used to build literacy in the general population and particularly among young people.

Why she is not optimistic, and what she wants Europe to do

On the geopolitics, Patrão Neves is blunt: the conditions for global cooperation do not exist, as the reaction of the President of the United States against the technology companies' own calls for moderation made clear. "Today the conditions are adverse" for a global solution, because there are "States competing with each other that think about their own lives and not about their populations or about humanity in general".

Her proposal is nonetheless a global one. "It would be important to seat industry and the major world leaders at the same table and attempt some form of pact, an elementary minimum governance." That first push at consensus, she believes, should come from the European Union. She accepts the initiative might not be well received, "but somebody has to take this step, and it has to be somebody with a clear awareness of the need for regulation throughout the world".

The Riyadh forum brought together governments, international organisations, academia, industry and civil society from more than 80 countries to discuss implementing UNESCO's 2021 Recommendation on the Ethics of AI. Portugal presented European projects alongside its own: Amália, the large language model built in Portugal; Mosaico, the common model for building digital public services; and an AI guide with an ethical risk assessment tool. Between 2013 and 2023, Portuguese public investment covered 410 AI-related research and development projects, which drew about 74.3 million euros from the Fundação para a Ciência e a Tecnologia.