by Dávid Szász
Artificial intelligence (hereinafter: AI) will change the world as we know it without doubt. For better or worse? That remains to be seen. The steps humanity will take in the forthcoming years in order to regulate the development and use of AI will decide the outcome. The future will be determined by our actions.
In the past few months several governments, international organizations, scholars, and academics have expressed their view on AI and made recommendations regarding the dangers it possesses. As a flagship organization of human rights protection in Europe the Council of Europe has already made and will make several significant steps in order to ensure and guarantee the undisturbed enjoyment of human rights all across its member states by mitigating the dangers presented by AI. In order to achieve this goal several bodies of the organization have been set into motion already to address the risks presented by AI to human rights and democracy.
As so far:
- The Secretary General of the Council of Europe has identified AI as one of the organization’s key strategic priorities in the Digital Agenda for 2022-2025.[1]
- The Parliamentary Assembly of the Council of Europe whose mission is to uphold the shared values of human rights called on the Committee of Ministers to elaborate a legally binding instrument governing AI with specific principles based on the protection of human rights and democracy.[2]
- The Committee of Ministers, the Council’s statutory decision-making body, has decided to establish the Committee on Artificial Intelligence (hereinafter: CAI) with the purpose of drafting a framework Convention on Artificial Intelligence.[3]
- The CAI, which was tasked with the drafting of the framework convention on the development, design, and application of AI based its work on the recommendations[4] of its predecessor, the Ad Hoc Committee on Artificial Intelligence (hereinafter: CAHAI). The draft convention will be founded on the Council’s standards on human rights, democracy, and the rule of law.[5]
The CAI started working in 2022 and worked throughout 2023, which culminated in the current version of the “Draft Framework Convention on Artificial Intelligence, Human Rights, Democracy, and the Rule of Law” (hereinafter: Draft Convention or Draft) by December 2023, which document contains the outcomes of the second reading and will serve as the basis for the third and final reading.[6]
It is important to note however, that the current version[7] of the Draft Convention on AI may be altered several times and severely before adoption, as it will be put under negotiations. The current Draft Convention also reflects this, as some of the Articles (such as Scope) or even the Preamble have several different text proposals. The current goal is to finalize the draft by March 2024 and adopt it at the ministerial level in May 2024.[8]
In the current state of the Draft Convention[9], it consists of sixteen pages and eight chapters. As set forth in the Preamble, the goal of the Convention is to oppose the misuse of AI[10] and to create a globally applicable legal framework setting out common general principles and rules governing the design, development, use and decommissioning of these systems. It will aim to ensure that activities within the whole lifecycle of AI systems are consistent with human rights and will oblige each party to adopt or maintain appropriate legislative, administrative measures to give effect to the provisions set out.
As for the scope of applicability, the Draft proposes several different text alternatives. However, the most significant parts are almost the same in the different versions of the text. Namely, the parts which state that the Convention shall only be applied to activities within the lifecycle of AI systems that have the potential to interfere with human rights, democracy, and the rule of law, and will not be applied to activities without such potential.
According to the Draft, two general obligations will be set out in the Convention. Under these obligations the parties must ensure the protection of human rights and the integrity of democratic processes. To achieve this, they must adopt or maintain measures to ensure:
- that the activities of AI systems are compatible with obligations to protect human rights;
- that AI systems are not used to undermine the integrity, independence and effectiveness of democratic institutions and processes, including the principle of separation of powers, respect for judicial independence, and access to justice.
Beside the obligations already mentioned, the following general principles will be laid out in the Convention for the parties to implement regarding AI systems:
- the respect of human dignity and individual autonomy;
- transparency and oversight;
- accountability and responsibility;
- equality and non-discrimination;
- privacy and personal data protection;
- preservation of health and the environment;
- reliability and trust;
- and safe innovation.
According to the Draft, parties will be obliged to ensure the availability of accessible and effective remedies for violations of human rights resulting from the activities of AI systems. They must also establish that, where an AI system substantially informs or takes decisions impacting on human rights:
- effective procedural guarantees and safeguards are available to the persons affected;
- persons interacting with an AI system are notified that they are interacting with such a system rather than with a human.
Regarding risk and impact management, measures shall be taken for the identification, assessment, prevention and mitigation of risks and impacts to human rights, democracy and rule of law arising from the design, development, use and decommissioning of AI systems. This may require the implementation of mechanisms for a moratorium or ban in respect of certain uses of AI systems where such practices are considered incompatible with the respect of human rights, the functioning of democracy and the rule of law. The nature and severity of such measures will depend on the risk associated with the AI system in question and by this the Convention will incorporate the so-called risk-based approach.
Important to note that the parties will have to implement the Convention without discrimination on any ground and take due account of any specific needs and vulnerabilities in relation to respect of the rights of persons with disabilities and of children. The protection of whistleblowers must also be ensured according to the Draft.
Regarding the relationship with other legal instruments the Draft clearly states that nothing in the Convention shall be construed as limiting, derogating from, or otherwise affecting the human rights or other legal rights and obligations which may be guaranteed under the laws of a party or under any other international agreement to which it is a party.
The parties to the Convention will also be obliged to establish or designate one or more effective mechanisms to oversee compliance with the obligations. Such mechanisms must exercise their duties independently and impartially and they must have the necessary powers, expertise and resources to effectively fulfil their tasks of overseeing compliance.
As can be seen, this Convention holds the potential to become another important pillar of human rights protection in Europe, just as the Council’s previous achievements, and will safeguard these rights against the threat posed by AI, which will ultimately result in the safe development and use of AI in Europe.
[1] See: Council of Europe: Digital Agenda 2022-2025., 2022 .05. 04., 13-14. p., https://rm.coe.int/coe-digital-agenda-2022-2025-pro-eng-web/1680aa3e1b (last accessed: 2024.01.11.).
[2] Council of Europe: The Council of Europe & Artificial Intelligence, 2023., 5. p., https://rm.coe.int/brochure-artificial-intelligence-en-march-2023-print/1680aab8e6 (last accessed: 2024.01.11.).
[3] Ref. 2., 4. p.
[4] See: Ad Hoc Committee on Artificial Intelligence: Possible elements of a legal framework on artificial intelligence, based on the Council of Europe’s standards on human rights, democracy, and the rule of law, 2021. 12. 03., https://rm.coe.int/cahai-2021-09rev-elements/1680a6d90d (last accessed: 2024.01.11.).
[5] Ref. 2. 10. p.
[6] Digital Watch Observatory: Convention on AI and human rights (CoE process), 2024., https://dig.watch/processes/convention-on-ai-and-human-rights-council-of-europe-process (last accessed: 2024.01.11.).
[7] For the current version see: Digital Watch Observatory: Convention on AI and human rights (draft December 2023), 2024., https://dig.watch/resource/convention-on-ai-and-human-rights-draft-december-2023 (last accessed: 2024.01.11.).
[8] Ref. 6.
[9] See: https://rm.coe.int/cai-2023-28-draft-framework-convention/1680ade043 (last accessed: 2024.01.11.).
[10] The Draft Convention defines AI as a machine-based system that for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that may influence physical or virtual environments.
