There were benefits and also risks. Experts had been issuing warnings about the technology for years, although she noted that women experts or experts belonging to minorities were not as listened to, as the powerful companies themselves. Action was needed in order to ensure that these risks didn’t exacerbate inequalities and didn’t just keep the powerful powerful and the vulnerable vulnerable. The Council of Europe was currently doing this by negotiating an AI convention to safeguard human rights, democracy and the rule of law. These core values could be affected by reliance or a choice to rely on those systems. Ms SMUHA elaborated on a few myths about AI regulation, which needed repudiating because they stood in the way of effective AI regulation. “Human beings are biased as well. So, why should we single out this technology?”, she said. Human beings were biased, but AI systems operated at a scale and speed that was unprecedented. One push of a button in an AI system could ruin the life of an entire population, she warned. So, there was a difference which merited ex ante obligations. To the myth that one should only focus on very advanced machine learning and not all software, she shared the examples of the childcare benefit scandal in the Netherlands; Clearview AI; and Cambridge Analytica, which were more advanced systems. However, she said that the British post office scandal ruined the lives of many people and caused a great human tragedy, despite being very basic software which one would not call AI today. She said that the Convention’s scope had to be as broad as possible in terms of the systems and domains that it covered. To the myth that it was just a matter of time for AI systems to get better and more accurate, that it was a developing process which would one day be free of risks, she shared another example: facial recognition technology. It is not as accurate for women, for example, because of its biased data. She assumed that in a few years cameras everywhere with great accuracy would be able to determine that her face had Jewish origins, and asked if that was much better for privacy and democracy as the AI systems would seemingly then “be more accurate”. It was not about the development of the technology, but also the use. The Convention needed to have clear red lines on what was acceptable and what was not. To the penultimate myth, about innovation being killed if AI were regulated, which would be detrimental, she had a question: Why caring about innovation unless it could enhance individual and societal welfare, as innovation could help in prosperity and promoting human rights? It made perfect sense for her that in the AI Convention being negotiated, boundaries in which innovation had to flourish needed to be shared and clarified, and it made sense to say it was democracy, human rights and the rule of law. The final myth was about it being too early to regulate because there were so many effects of AI that we did not know about yet. She pointed out that we did know that there was harm to human rights, democracy and the rule of law, as seen in private uses of AI and in public uses of AI. It was very important to protect those rights and values, because the harm could not only be on a huge scale, but it could also be irreversible. It was not too early to do something about it, she concluded. Ms BRYNJÓLFSDÓTTIR said these were very interesting points. She said Ms Tinna HALLGRIMSDÓTTIR, youth representative in the Icelandic Climate Council, had been very active when it came to being Chair of the Icelandic Youth Environmentalist Association, and raising the current and the future challenges of the climate crisis and environmental issues. She asked whether Europe's leaders were doing the right things to identify, protect, and adapt to these emerging new human rights, particularly when it came to the right to a clean, healthy and sustainable environment. Ms Tinna HALLGRIMSDÓTTIR, youth representative in the Icelandic Climate Council, said the simple answer to the question was, no, they were not doing enough, not reacting well enough. A more complex question was how they should frame the issues and how that could contribute to the problem. Ms HALLGRIMSDÓTTIR said she felt like they were falling into a trap of framing the right to a clean, healthy and sustainable environment as merely a right for future generations, or merely as a new right. She wanted to emphasise that the right to a clean, healthy and sustainable environment was not only a right for future generations, but it was as much of a necessity for each and every one of them sitting there today.

20 DE JULHO DE 2023 ____________________________________________________________________________________________________________

27