The Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law
Keynote speech by Director Hanne Juncher, Council of Europe
Thank you very much for the invitation to speak at this prestigious event, and in such impressive company.
- Background to the new Convention
- The content of the Convention
- What we see as the next steps.
AI comes with this great promise of innovation - in science, education, health care, business, climate, administration - and it’s already changed how we interact and communicate.
But, as we heard earlier, AI also comes with another side. For those in bad faith, it can be used to undermine democratic processes, including by spreading lies and fake news. It can be used for surveillance purposes – to control the population and marginalise political protest and activity.
It can skewer, and promote bias, in decision-making, and lead to discrimination and exclusion, and other breaches of human rights.
It’s not the technology per se that poses these problems of course. We shouldn’t fear the algorithms. But we should be wary of how, and in what contexts, they are being developed and used.
It’s with all this in mind that the Council of Europe has drawn up a new international treaty - the world’s first I’m proud to say: the Framework Convention on AI and Human Rights, Democracy and the Rule of Law. We did this in close cooperation with our like-minded partners in Europe and beyond.
The Council of Europe began by setting up an inter-governmental Committee to examine the need and prospect of actually regulating AI in international law. That resulted in a feasibility study which concluded that there was such a need.
That was followed by an analysis to outline the main elements that might be included in a legally binding treaty on AI, with the focus on human rights, democracy and the rule of law.
Then, a new Committee was set up by the governing body of the Council of Europe, which is the Committee of Ministers, and that was the Committee on AI, or the “CAI”.
The CAI was instructed by the Committee of Ministers to establish an international negotiation process for the purpose of elaborating a legally binding treaty on AI, which would have global reach and be conducive to innovation.
The CAI elected as its Chair Ambassador Thomas Schneider from Switzerland who is with us today. He led the entire process and is speaking in the panel later today.
The CAI Committee successfully finalised the negotiations of the new treaty in March of this year.
In addition to the member states of the Council of Europe, we had Argentina, Australia, Canada, Costa Rica, the Holy See, Israel, Japan, Mexico, Peru, the United States, and Uruguay participating in the negotiations as observers. The European Union negotiated for its 27 member states.
We benefited from the input of 68 representatives of civil society, industry, academia, and international and regional organisations, who also contributed to shaping the new treaty. This helped make the process as participatory and inclusive as possible.
I should add that the Convention in its final form is fully compatible with the EU’s AI Act.
The final instrument was adopted by the Committee of Ministers in May, and it was opened for signature in Vilnius in Lithuania, in September.
On that occasion, the Convention was signed by 9 countries: Andorra, Georgia, Iceland, Norway, the Republic of Moldova, San Marino, the United Kingdom, Israel, and the United States. In addition, the European Union signed on behalf of its 27 member states.
It’s our hope that this – again, the world’s first international treaty on AI - will become the global standard for regulating the development and use of AI systems in contexts that are human rights, democracy and rule of law-sensitive.
The principles set out in the Framework Convention are applicable to both public and private sector AI systems.
Those principles include:
- respect for human dignity and individual autonomy
- transparency and oversight
- accountability and responsibility
- equality and non-discrimination
- privacy and personal data protection
- reliability, and
- safe innovation.
It’s worth noting that the Framework Convention doesn’t create new human rights specific to AI, but it ensures that all existing human rights will prevail, also in situations where an AI system is involved in taking, or informing, decisions.
Importantly, the Convention doesn’t apply to trivial AIs, but only comes into play for AI solutions that may pose a risk to human rights, democracy or the rule of law.
The Convention requires the parties to provide remedies and procedural safeguards, and to put in place a risk and impact management framework.
The CAI Committee is currently developing a methodology which Parties may use for the risk and impact management. It’s called the HUDERIA and it will be ready in a few months. Unlike the Convention, it will be non-legally binding.
Looking more generally, it seems to us that the challenges that AI technology pose to society can’t be addressed in any real sense only at the national, or even regional, levels. To be effective, there has to be a global legal framework, creating a level playing field for all the relevant actors, across regions, and providing for international cooperation. This is what the Council of Europe Convention offers.
We still hear the argument that regulation is a hindrance for technological innovation and the resulting economic growth. We don’t believe that this is the case, provided that the regulation is properly targeted and calibrated.
In fact, such regulation may even improve the chances of economic growth. Because it favours responsible innovation, and it allows societies to make use of the technology without having to fear unintended consequences in the form of abuse and misuse by the state, or by a private entity.
It also allows for the level legal playing field that I mentioned earlier in respect of companies working globally.
We basically believe that new and emerging digital technologies, such as AI, will prove to be of benefit for humanity. AI technologies offer prospects of economic and scientific development to societies across the world. But if they’re used in an irresponsible, or even deliberately harmful, way, then we risk seeing human rights, democratic processes, and the rule of law undermined.
The new Convention offers a very good framework for reducing those risks, ensuring transparency and providing some concrete reassurances to those who will experience the actual consequences of AI solutions.
At the same time, the way we see it, there is still a very uneven access to AI technology across the world’s regions. And uneven protection against abuses and misuse.
That makes it even more important to react now. With the Framework Convention we are, hopefully, contributing to bridging that divide.
I want to close by simply encouraging all like-minded states, members of the Council of Europe and non-member states, to accede to the new Convention.
And thank you again for this excellent event.