Civil society responds to the Council of Europe Treaty on AI
Together with other observer civil society organizations in the Committee on AI in the Council of Europe, AlgorithmWatch Switzerland stresses the importance of that legal framework on AI based on human rights, democracy, and the rule of law that is currently being elaborated in Strasbourg. We urge the EU not to delay this process in light of the negotiations on its own AI Act currently ongoing in Brussels. The two frameworks have a different purpose and should complement rather than copy-paste each other.
Civil Society organizations would like to respond to the EU statement directed toward the members of the Council of Europe (CoE) Committee on AI (CAI) and Observers (which includes Civil Society Organizations) regarding the upcoming/draft COE Convention on Artificial Intelligence.
We begin by noting that Artificial Intelligence is a rapidly evolving technology that touches every aspect of our lives. These systems make decisions every day about who is hired, who receives public benefits, who is admitted to university, and who is sentenced to prison. Few technologies have permeated our lives so quickly or raised so many questions about human dignity, autonomy, and justice.
As Civil Society Organizations, we have worked to document these problems and to propose solutions. We share the views of others that new technologies can be a force for good and advance social progress. But this necessarily requires the democratic rules to ensure that we control the use of these technologies, that they do not control us. The establishment of rules for governance for AI is central to this work.
Civil society closely follows the work of the European Union to develop a comprehensive regulatory framework for AI as we strongly support the work of this Committee to prepare a Convention on AI based on the Council of Europe and its mandate. We actively participate in the work of the EU institutions on the AI Act. We have also devoted two years to the work of the CoE AI preparatory committee, the CAHAI. We believe the final report of the CAHAI sets out many of the legal elements that are necessary for a comprehensive AI treaty to safeguard fundamental rights, protect democratic institutions, and uphold the rule of law.
We are therefore surprised and disappointed to see the Statement issued by the European Union regarding this undertaking, which essentially suspends the participation of EU states in the CAI process – states that have already actively contributed to the CAHAI recommendation – while the EU sorts out its participation in this matter.
To be clear, there are currently areas of overlap between the proposed Artificial Intelligence Act of the EU and the Convention on Artificial Intelligence of the CoE. But it is also clear that the EU and the CoE have different mandates, different constituencies, and serve different purposes. The ministries participating in the work on the EU AI Act primarily focused on the regulation of the internal market. The ministries participating in the work on the COE AI Treaty are responsible for the protection of fundamental rights.
We fully understand the need for a coordinated approach. At the same time, we see no need to suspend the forward progress of the work of the Committee on AI. We are further concerned about certain suggestions that appear to invite additional delay in the CoE process and to limit the outcome of the CoE’s work.
- We believe the proposal to focus on provisions of lesser consequences, such as the preamble, will delay forward progress on the CoE legal instrument. Many of the difficult issues were already addressed in the final report of the CAHAI, and received support from the EU member states. The Committee on AI should move forward with the discussion and debate on the draft text seriatim as is customary for all such working parties.
- We also object to the EU’s suggestion that there is a need to substantially revise the communications and consultation process. We wish to make clear our strong support for the inclusion and transparency procedures previously adopted by the Council of Europe to enable civil society and other observers to participate in the drafting and development of the AI Convention.
- The proposal of the EU to “align” the work of the CoE CAI with that of the EU is also of great concern. The EU AI Act contemplates a risk-based framework for the regulation of certain AI systems while the CoE anticipates an international treaty based on fundamental rights, democratic institutions, and the rule of law. Both approaches are important and necessary, but it would appear institutionally impossible to align these two approaches without diminishing the role of the CoE.A CoE Convention also creates an opportunity for non-CoE member states around the world to adopt the convention in the future. As such the impact of the convention to protect fundamental rights, rule of law, and democratic institutions can be magnified. A delayed process and/or a weakened convention would be a setback globally.
- The negotiations process within the EU is dynamic and further delays are possible. If the Commission is unable to negotiate on behalf of the EU prior to completion of the trilogue, this would massively delay the process of the CoE to protect fundamental rights, democracy, and the rule of law regarding the use of AI systems – within the EU but also beyond, since almost half of the CoE member states are not EU member states. It would thus also illegitimately put a halt to the efforts of many non-EU member states, which will not be subject to the AI Act.
It is vitally important to move forward the CoE work on Artificial Intelligence. We wish to remind the EU that other international organizations, recognizing the urgency of this moment, produced substantial frameworks for Artificial Intelligence in less two years.
- The OECD launched its work on AI in 2017 and set out the OECD AI Principles, with the support of 42 countries in May 2019. And substantial work on implementation is already underw
- The G20, representing the world’s largest economies, endorsed the OECD AI Principles as their own in June 2021 and continue to promote the Guidelines in their national AI strategies.
- UNESCO launched its formal work on AI 2019. A final Recommendation, with the support of 193 nations, was adopted in 2021. And that was a sweeping framework for AI that touches upon many topics, such as Climate Change, Sustainability, and Gender Equity, that are not even a part of this work.
More than two years have passed since the Council of Europe launched its work on Artificial Intelligence. And two years have passed since the EU issued its White Paper. It is time now to move toward the finish line.
We urge the EU to consider carefully the consequences of delaying the important work of the COE Committee on AI.
There is no pause button for the deployment of AI systems. If anything, AI holds down the button labeled “fast forward.” This is not the moment to hit the pause button on democratic governance/policy-makinginstitutions. The work of the Council of Europe on the AI Treaty must continue without undue delay.
Endorsements
AlgorithmWatch
AlgorithmWatch Switzerland
Center for AI and Digital Policy
Conference of INGOs of the Council of Europe
Fair Trials
Global Partner Digital
Homo Digitalis
Istanbul Bar Association
Read more on our policy & advocacy work on the Council of Europe.