A new legal standard for AI
Countries worldwide have enacted laws to regulate AI, often targeting specific issues like privacy, misinformation and its applications – think of online recommendation systems. However, the AI Framework Convention is the first legally binding international treaty covering all AI applications, valid for all countries that join. In addition to the 46 member states of the Council of Europe, several countries from different regions participated in the negotiations – including Argentina, Australia, Canada, Costa Rica, the Holy See, Israel, Japan, Mexico, Peru, Uruguay and the United States.
The Framework Convention provides a legal structure for the entire lifecycle of AI systems. It promotes innovation in the field of AI, while mitigating risks to human rights, democracy and the rule of law.
National laws must not contradict international treaties. If necessary, governments must amend their national laws in advance of, or at the same time as, the ratification (approval) of the treaty. 'The EU has announced that it will implement the Framework Convention through existing laws such as the AI Act and, probably, the GDPR and the DSA,' says Silvia De Conca. This is remarkable, as the AI Act primarily focuses on market safety and does not offer direct solutions to citizens harmed by AI systems.
European legislation and the Framework Convention: conflict or synergy?
De Conca investigates how EU AI legislation aligns with the Framework Convention, and whether the convention provides sufficient protection for citizens. 'Given that the AI Act and the Framework Convention influence each other, I don’t see any major legal conflicts,' she says. 'But there are challenges. The Framework Convention expects states to actively protect citizens and offer the possibility for them to file complaints with appropriate judicial or administrative authorities, if that protection falls short. This could cause friction.'
'The AI Act does not currently regulate compensation for violations of fundamental rights. Facial recognition technologies, for example, especially when used remotely and in real-time in public spaces, can infringe on citizens' freedom of expression and association. They may discourage peaceful protests and the exercise of democratic rights. At the same time, they can be used to track criminals or locate missing persons, 'De Conca explains. 'Legal solutions, such as filing objections or accessing courts, depend on national legislation – which varies widely across EU countries. This can clash with the Framework Convention’s goal of ensuring a uniform and substantial level of protection.'
The importance of coherent AI legislation
AI regulation requires strong coordination between national and international laws, according to De Conca. 'The EU has many ambitious laws to regulate digital technology, which is positive. But they must align well with each other to avoid implementation issues.'
She compares legislation to a living organism: 'Laws grow and change with society. But if one organ does not work in harmony with the rest, the whole system suffers.'
Ethical issues in AI and big data
Beyond legal aspects, AI also raises profound ethical dilemmas. Facial recognition technology, for example, can help find criminals or missing persons, but may also suppress freedom of expression and the right to protest. 'Where do we find the balance between security and civil rights?' De Conca asks.
Moreover, the use of AI and big data exacerbate global inequalities. 'AI requires vast amounts of resources: water, land, minerals and electricity. Countries in the Global South often bear the burden of technology production, such as resource extraction and environmental and labour impacts, while the benefits are concentrated elsewhere,' says De Conca. 'This intensifies existing inequalities: the rich get richer, while the poor get poorer.'
Pact for the Future
Safe AI systems, ethical concerns and inequality are also high on the United Nations’ (UN) agenda. Through the Pact for the Future and its annex, the Global Digital Compact, the UN aims to bridge the digital divide and support developing countries in implementing technological solutions and safe AI systems.
'It’s crucial that the UN takes on this role. Without them, there would be no global platform to discuss these issues,' emphasises De Conca. 'Even if progress is sometimes slow, the impact is invaluable.'
In summary
The Council of Europe's Framework Convention provides an essential legal foundation for balancing technological innovation and human rights, proposes Silvia De Conca. However, the key question remains whether existing EU laws, such as the AI Act, offer sufficient protection. According to De Conca, the coming years will be crucial in charting the right course and ensuring that AI serves society, without endangering fundamental rights.