Earlier this month, the Council of Europe adopted a new legally binding international agreement on AI called the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law that was signed by the EU, the UK, Norway, the US, and others. This convention is the first legally binding multilateral instrument governing artificial intelligence and is intended to create a legal framework for ensuring that all activities within the AI lifecycle respect human rights, democracy, and the rule of law.
As we all know by now, AI has the potential to have numerous legal implications in regard to privacy, intellectual property, user safety, and discrimination. This new Council of Europe AI Convention attempts to set legal obligations on states that join this convention to ensure that all activities within the lifecycle of artificial intelligence systems are consistent with human rights, democracy and the rule of law, both in regard to activities undertaken by public authorities as well as private actors.
The convention sets key principles that states need to respect, such as human dignity and individual autonomy, transparency and oversight, accountability and responsibility, equality and non-discrimination, privacy and personal data protection, reliability, and safe innovation. The convention requires states to introduce measures that require that persons affected by AI systems are notified of the use of AI and grant such persons the right to challenge decisions relating to them that were made by AI systems. The convention also requires states to have measures that require carrying out AI impact assessment on human rights, democracy, and the rule of law.
The convention does not have many specifically prescribed actions that states need to take, and appears to be written in a manner that gives states more flexibility in regard to the measures that states are required to impose against private actors. It also gives states the ability to introduce provisions in a gradual manner while taking into consideration the unique context of each country.
The convention does not appear to have teeth, as there are no serious consequences for states that fail to meet their obligations. There is also no formal complaints procedure that allows individuals to take action against states that violate their obligations. Furthermore, the convention excludes from its scope activities within the lifecycle of AI systems relating to the protection of national security interests, and there is a separate paragraph that states that “matters relating to national defence do not fall within the scope of this Convention”. This is disappointing as it means that the convention will probably not be useful to stop states such as Israel from using AI systems, like Gospel and Lavender, in military operations, even if these acts are deemed war crimes by the ICJ.
It is also worth noting that this convention is an international agreement that is only binding on the countries that choose to sign it and then complete their domestic constitutional formalities to ratify it. At the time of writing, it has been signed by 10 states plus the EU. However, the treaty will only enter into force after at least 5 states that sign it complete their domestic ratification processes. This treaty is also not open for Council of Europe non-members to join, unless they are invited.
Oman has never joined any Council of Europe convention in its history, with the exception of the Convention on Mutual Administrative Assistance in Tax Matters which was a joint project by the Council of Europe and the OECD. It is practically impossible that Oman would ever join this convention.
However, in the same way the EU AI Act that was adopted earlier this year is likely to be used as a benchmark for any new AI laws issued in Oman or elsewhere, it is not unlikely for the GCC or the Arab League to decide to take part in this new trend of AI regulation and develop their own AI treaty using the same principles that the Council of Europe has now codified as principles of public international law.
The Council of Europe AI Convention will only enter into force after the ratification of 5 member states from among those who sign the convention. You can read it in full on this link.
