• The United States, the European Union, and the United Kingdom are anticipated to sign the Council of Europe’s new convention on artificial intelligence (AI), marking it as the first “legally binding” international treaty on AI.
• The treaty, officially named the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, was opened for signature during a Council of Europe Ministers of Justice conference in Vilnius, Lithuania.

Purpose and Approach:

• Human Rights Focus:
  • The convention prioritizes human rights in regulating both public and private-sector AI systems.
  • It seeks to address concerns that fragmented regulations by individual countries could impede the development of AI technology.
• Risk-Based Framework:
  • The treaty adopts a risk-based approach to the design, development, use, and decommissioning of AI systems.
  • It is applicable to AI systems used across public sectors, including entities acting on behalf of the government, and private sectors, covering multiple geographies.

Key Provisions and Obligations:

• Accountability:
  • Signatories are accountable for harmful and discriminatory outcomes resulting from AI systems.
  • They must ensure AI outputs respect equality and privacy rights and provide legal recourse for victims of AI-related rights violations.
• Human Rights and Democratic Integrity:
  • AI systems must be consistent with obligations to protect human rights.
  • They must not undermine democratic institutions, including judicial independence and access to justice.
  • Measures must be implemented to safeguard democratic processes, including fair access to public debate and freedom of opinion.
• Exemptions:
  • Exemptions include national security and research and development activities.
  • Parties are obligated to address risks related to the lifecycle of AI activities by both public and private actors.

International Context:

• Related Agreements:
  • The treaty complements other AI regulations and agreements, such as the G7 pact on AI (October 2023), Europe’s AI Act, and the Bletchley Declaration signed by 28 countries (November 2023).

Challenges and Concerns:

• Enforcement Issues:
  • Despite being termed “legally binding,” the treaty lacks specific provisions for punitive sanctions like penalties or fines.
  • Compliance is primarily monitored, which may not be a strong deterrent for enforcement.
• Initial Approvals:
  • Hanne Juncher, Director of Security, Integrity and Rule of Law at the Council, indicated that 10 participants are expected to be among the first to approve the convention.
  • This early support reflects the significant investment and satisfaction of these participants in the treaty’s outcome.

Conclusion:

• The Council of Europe’s convention represents a pioneering step towards a globally coordinated approach to AI regulation, emphasizing human rights and democratic principles. However, its effectiveness will depend on the development of robust enforcement mechanisms and broader adoption by key international stakeholders