Context

The global governance of Artificial Intelligence (AI) is becoming increasingly intricate as nations strive to manage AI within their borders through various legislative measures and executive orders. While numerous experts have suggested the need for a global AI treaty, the obstacles to achieving such a comprehensive agreement are significant. Amid this complexity, the Council of Europe (COE) has taken a notable step by adopting the Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, commonly referred to as the ‘AI convention’. This development marks a significant milestone in AI governance, linking it directly to human rights, democracy, and the rule of law.

Europe’s AI Convention

Despite the plethora of ethical guidelines, ‘soft law’ tools, and governance principles articulated in various documents, none of these are binding nor appear likely to culminate in a global treaty. Moreover, there are currently no active negotiations for an AI treaty at the global level. Against this backdrop, the COE's adoption of the AI convention on May 17 represents a significant advancement. Established in 1949, the COE is an intergovernmental organization with 46 member states. The AI convention is set to open for signature on September 5 and aims to provide a comprehensive framework for AI governance, emphasizing its alignment with human rights, democracy, and responsible use.

What is a Framework Convention?

A ‘framework convention’ is a type of legally binding treaty that outlines broad commitments and objectives. It establishes mechanisms for achieving these goals, while specific targets and measures are defined in subsequent agreements known as protocols. For instance, the Convention on Biological Diversity is a framework convention, and the Cartagena Protocol on Biosafety is a protocol under it, addressing issues related to living modified organisms.

The utility of the framework convention approach lies in its flexibility. It allows the core principles and processes necessary to achieve the convention's objectives to be encoded, while granting parties the discretion to decide how to implement these principles based on their individual capacities and priorities. The AI convention, therefore, can serve as a catalyst for similar conventions at regional levels elsewhere. Additionally, as the United States is a COE member, the AI convention could indirectly influence AI governance in the U.S., a key player in global AI innovation.

Scope of the Convention

The AI convention’s scope is defined in its articles. Article 1 asserts that the convention aims to ensure that all activities within the AI system lifecycle are consistent with human rights, democracy, and the rule of law. Article 3 expands on this by stating that the convention covers activities within the AI system lifecycle that could potentially interfere with these principles. It mandates that each party applies the convention to activities by public authorities or private actors acting on their behalf. Furthermore, it requires parties to address risks and impacts arising from private sector AI activities in a manner that aligns with the convention's objectives.

This broad scope ensures comprehensive oversight of AI systems, aiming to mitigate any negative impacts on fundamental rights and democratic processes. By embedding these requirements, the convention seeks to create a governance structure that prioritizes human rights and ethical considerations throughout the AI lifecycle.

Addressing National Security Concerns

The AI convention also addresses national security concerns, although with notable exemptions. Articles 3.2, 3.3, and 3.4 outline these exemptions, which pertain to protecting national security interests, research, development and testing, and national defense. As such, military applications of AI are excluded from the convention’s purview. While this exclusion might raise concerns, it is a pragmatic move, acknowledging the current lack of consensus on regulating military AI applications.

However, these exemptions are not absolute. The convention does not entirely rule out its applicability in relation to national security and testing. This nuanced approach allows the convention to address a broad range of AI activities while recognizing the complexities and sensitivities associated with national security.

General Obligations and Implementation

The convention imposes several general obligations on its parties. Article 4 emphasizes the protection of human rights, while Article 5 focuses on the integrity of democratic processes and respect for the rule of law. Although issues like disinformation and deep fakes are not explicitly mentioned, Article 5 expects parties to take steps against such threats, reinforcing the convention’s comprehensive approach to AI governance.

The convention also highlights the need for effective remedies and procedural safeguards. Article 14 mandates that governments install measures to address grievances related to AI system impacts, while Article 15 requires procedural safeguards to ensure compliance with the convention’s objectives. These provisions underscore the convention's commitment to maintaining robust oversight and accountability mechanisms.

Why We Need the AI Convention

The AI convention does not introduce new human rights specific to AI. Instead, it reaffirms that existing human and fundamental rights, as protected by international and national laws, must remain safeguarded during the application of AI systems. This reaffirmation is crucial, as it ensures that AI deployment does not erode established human rights frameworks.

The primary obligations under the convention are directed towards governments, who are expected to implement effective remedies and safeguards. By taking a comprehensive approach, the convention aims to mitigate risks associated with AI systems, protecting human rights and democratic processes. While challenges to its implementation are inevitable, especially given the rapid pace of technological advancements outstripping policy development, the convention represents a significant step towards responsible AI governance.

Conclusion

In conclusion, the COE’s adoption of the AI convention marks a significant advancement in the global governance of AI. By linking AI governance to human rights, democracy, and the rule of law, the convention sets a comprehensive framework that addresses the ethical and societal impacts of AI systems. Although challenges to its implementation remain, the AI convention's framework provides a flexible and robust approach to ensuring that AI development and deployment align with fundamental rights and democratic values. As nations continue to navigate the complexities of AI governance, the AI convention serves as a critical tool for fostering responsible and ethical AI practices worldwide.

Source - The Hindu