Ethics and AI | Part 5

No man is above the law, why should be AI?

The Framework Convention on Artificial Intelligence and human rights, democracy, and the rule of law

Since 2024, the international community could be satisfied that there is no longer a legal vacuum on AI. The Council of Europe adopted a Framework Convention on Artificial Intelligence and human rights, democracy, and the rule of law¹, which is the first-ever international legally binding treaty in this field. As the title of the Convention indicates, the Council of Europe places its document in a very clear context and ambit, one that fits its mandate and competence as an organisation: human rights, democracy, and the rule of law. Therefore, it is reasonable to expect that while it remains technology-neutral, the Convention will not regulate technology but the way the human factors will use it. So, the first reading indicated that the Convention came as close as it could to human responsibilities as well as to the meaning of international cooperation on artificial intelligence.

The essential meaning of the convention is candidly explained in two preamble paragraphs, which are a paradoxical juxtaposition that say everything and its opposite:

Recognizing that activities within the lifecycle of artificial intelligence systems may offer unprecedented opportunities to protect and promote human rights, democracy and the rule of law;

Concerned that certain activities within the lifecycle of artificial intelligence systems may undermine human dignity and individual autonomy, human rights, democracy and the rule of law;

The solution to this antithetical enouncement goes as easy as it comes, at least in words, in the operative part, Article 1 (1) of the Convention:

The provisions of this Convention aim to ensure that activities within the lifecycle of artificial intelligence systems are fully consistent with human rights, democracy and the rule of law.

Since 2024, the international community could be satisfied that there is no longer legal vacuum on AI.

An obvious merit of the Convention for the researcher is that it offers a definition of artificial intelligence system, which would spare the author and the readers of the effort to consult a literature abundant in personal perspectives and often biased and inconsistent. Article 2:

For the purposes of this Convention, “artificial intelligence system” means 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. Different artificial intelligence systems vary in their levels of autonomy and adaptiveness after deployment.

Very importantly, the Convention implies the responsibility of both public authorities or private actors acting on their behalf. (Article 3a)

As we may understand private actors that are not acting on behalf of the public authorities, are not concerned by the Convention. Or, how many companies that produce AI systems, spread or sell them, are doing it on behalf of the public authorities? The drafters of the Convention were very cautious about intruding on the interests of the industry. The Convention does not apply to “activities within the lifecycle of artificial intelligence systems related to the protection of its national security interests”, nor “to research and development activities regarding artificial intelligence systems not yet made available for use. (Article 3.2 and 3.3.).

With these areas and actors taken out of the purview of the Convention, what is left for regulation? The Convention specifies two general obligations: protection of human rights (Article 4) and maintaining the integrity of democratic processes and respect for the rule of law (Article 5). Yet, these obligations are already part of a solid corpus of international law, and they have an erga omnes vocation. The Convention does not bring anything new, but just state the obvious. The notion of artificial intelligence system can be replaced with blockchain applications, drones, satellite systems, or cloud seeding devices.

The principles stipulated by the Convention do not come with anything that would deal with issues which we have identified as fundamental in defining ethics: good and bad, right or wrong etc., pertaining to AI forms of manifestation. Those principles (human dignity and individual autonomy, transparency and oversight, accountability and responsibility, equality and nondiscrimination) have to be respected “in relation to activities within the lifecycle of artificial intelligence system”. As some of these principles are already embedded in language of hard international law, the Convention, as praiseworthy as it may be, does not bring new lights and does not address the main existential and epistemological worries, or about the human condition in general. The risk management measures provided by the Convention (Article 16) are abstract and vague, and circumscribed by words of caution: “graduated”, “differentiated”, “where appropriate”.

The only reference to ethics appears only once in the Convention, which recommends public discussion and multistakeholder consultation on implications of the artificial intelligence systems, including the ethical ones. (Article 19).

We should therefore look for more contents, in other attempts to deal with ethics.

[Part 5 of 6-part series]

---

1. Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, Council of Europe Treaty Series – No. 225, 5.IX.2024 ↩︎

---

Dr Petru Dumitriu was a member of the Joint Inspection Unit (JIU) of the UN system and former ambassador of the Council of Europe to the United Nations Office at Geneva. He is the author of the JIU reports on ‘Knowledge Management in the United Nations System’, ‘The United Nations – Private Sector Partnership Arrangements in the Context of the 2030 Agenda’, ‘Strengthening Policy Research Uptake’, “Cloud Computing in the United Nations System”, and “Policies and Platforms in Support of Learning”. He received the Knowledge Management Award in 2017 and the Sustainable Development Award in 2019 for his reports. He is also the author of the Multilateral Diplomacy online course at DiploFoundation.