AI is changing the world, but how can we ensure we develop and use it ethically?
In September 2024, the new Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law was opened for signature by the Council of Europe (CoE) – also known as the name of AI Convention or AI Treaty.
Following in the footsteps of the European AI law adopted earlier this year, the AI Convention is the first internationalone-of-a-kind framework covering the ethical development and use of AI.
Reach and participation in relation to EU AI law
The AI Convention still shares some fundamental principles with EU AI law, but has some fundamental differences:
Specific provisions
| European AI law | AI Convention |
| Covers detailed provisions on the development and use of the GPAI (General Purpose AI) system | Focuses more on broader principles to ensure that human rights, democracy and the rule of law are not affected – it does not specifically address GPAI systems |
Scope and competence
| European AI law | AI Convention |
| Applies only to EU member states, with emphasis on market regulation, product safety and consumer protection. It details specific risk categories with obligations by level of risk, i.e. unacceptable, high, limited or minimal. | Potentially global in scope, open for signature by the Council of Europe and non-member countries. It uses a risk-based approach, but does not establish specific risk categories for AI systems; rather, the risk is assessed against the potential impact of AI on human rights, democracy and the rule of law. |
Legal nature
| European AI law | AI Convention |
| Detailed and prescriptive regulations with specific requirements and prohibitions to follow if you operate within the EU. It has specific enforcement mechanisms and sanctions in the event of non-compliance. | Establishes general commitments regarding the use, risks and potential impact of the AI system. It leaves specific implementation and compliance details to national legislation aligned with the Framework Convention, with the possibility of sanctions, bans or other independently decided measures if the use of AI enters into in conflict with the agreed arrangements. |
What else does the Framework Convention on AI cover?
The AI Convention aims to enable broader international adoption and cooperation around the regulation of AI. Although these requirements may be considered less stringent than EU AI law, signatories to the framework indicate their commitment to:
- Prioritize human rights, including respect for human dignity, non-discrimination, data protection and privacy in the use and development of AI
- Transparency requirements for AI-generated content and disclosure of interactions with AI systems
- A risk-based approach to the regulation of AI, in particular on the risks and potential impact of AI systems on human rights, democracy and the rule of law
- A focus on AI supervised by humans, with a focus on transparency, accountability, documentation of use and monitoring
- Principles for reliable AI that prioritizes data security and governance
- Documentary obligations for high-risk AI systems
- Support for safe innovation via regulatory sandboxes
Another important element of the AI Convention is that it provides an additional degree of flexibility to private sectors, giving them the possibility to apply the obligations set out in the AI Convention or to implement other appropriate measures.
It also includes certain exemptions for research and development and activities involving national security – provided that human rights obligations still apply.
When will the Framework Convention enter into force?
The Framework Convention on AI has already garnered signatures from key players such as the United States, Canada, Japan, Australia and the United Kingdom – and the door remains open for other countries to that they join efforts to shape a responsible AI future.
However, the AI Convention requires specific conditions to be met before it can enter into force:
- At least three of the signatories must be member states of the Council of Europe.
- It needs at least five signatories to ratify it. This is a separate process that usually involves national legislative approval.
- A waiting period of three months after the conditions are met before the entry into force of the treaty, the first of the month following the expiration of this period
At the time of writing this article, in October 2024, the current signatories to this framework had not yet completed this ratification process.
Review and next steps
While this landmark treaty highlights the growing global focus on ethical AI, concerns have been raised about the applicability of the AI Convention and potential gaps therein.
In a recent article in ReutersFrancesca Fanucci, a legal expert at the European Center for Nonprofit Law (ECNL) who was involved in the treaty drafting process, warned that the agreement’s principles are so broad that they could be difficult to apply in practice. She highlights national security exemptions and limited oversight of private companies as particular areas where double standards are of concern.
It remains to be seen how these concerns will be addressed. At the same time, Council of Europe Secretary General Marija Pejčinović Burić urged more countries to sign the AI Convention, urging those who have already signed it to ratify it as soon as possible so that it can come into force.
The content of this article is intended to provide a general guide on the subject. Specialist advice should be sought regarding your specific situation.
