Need for International Governance of Artificial Intelligence

UN Secretary-General António Guterres established the High-level Advisory Body on Artificial Intelligence “to undertake analysis and advance recommendations for the international governance of artificial intelligence” in October 2023. Facing human rights violations in the field of, by, or with artificial intelligence as well as the potentially catastrophic consequences of unchecked artificial intelligence, the need for international governance of artificial intelligence is broadly recognized. At the launch of the advisory body, UN Secretary-General António Guterres acknowledged the positive potential of Artificial Intelligence while simultaneously warning that: “As things stand, AI expertise is concentrated in a handful of companies and countries. This could deepen global inequalities and turn digital divides into chasms. The potential harms of AI extend to serious concerns over misinformation and disinformation; the entrenching of bias and discrimination; surveillance and invasion of privacy; fraud, and other violations of human rights.”

As required, the Advisory Body submitted an interim report in December 2023. The Advisory Body is mandated, “based on feedback to its interim report”, to submit a second report by 31 August 2024 “which may provide detailed recommendations on the functions, form, and timelines for a new international agency for the governance of artificial intelligence”.

I have reviewed the interim report by the High-level Advisory Body on Artificial Intelligence (co-chairs: Carme Artigas, Secretary of State for Digitalisation and Artificial Intelligence of Spain; James Manyika, Senior Vice President of Google-Alphabet) to present the following comments on the contents.

‘Co-Existence’ Report Instead of ‘Protection Report’

The Interim Report of the Advisory Body is essentially an exploratory one, discussing relevant principles and norms, and touching on possible institutional functions for a not yet delineated future body. It does not yet address “the functions, form, and timelines for a new international agency for the governance of artificial intelligence.”

The Interim Report may be deemed a ‘co-existence’ report (co-existence between humans and artificial intelligence), whereas what is be required is a ‘protection report’ (protecting humans from artificial intelligence). This perspective is fundamental to future work in this area.

In the annual Romanes lecture at Oxford University in February 2024, Geoffrey Hinton, a leading authority in this field, spoke to the question “Will Digital Intelligence Replace Biological Intelligence?” and answered in the affirmative: ‘almost certainly, yes.’ He told a Financial Times interviewer before his lecture: “If I were advising governments, I would say that there’s a 10 per cent chance these things will wipe out humanity in the next 20 years. I think that would be a reasonable number”. Geoffrey Hinton warns that Artificial Intelligence could become more intelligent than humans including the power to decide about the future of humans also in a negative way. There is debate on Hinton’s assessment, but the question to be asked is whether one can take it lightly in view of his experience and eminence in the field. It has become clear that future deliberations on this issue must give priority to the issue of the protection of humans from Artificial Intelligence guaranteeing all humans human rights and fostering the sustainable flourishing of all humans and the planet.

The UN Secretary-General initially framed the mandate of the Advisory Body in terms of the international governance of artificial intelligence. He has asked the Board for detailed recommendations on the functions, form, and timelines for a new international agency for the governance of artificial intelligence. The Advisory Body must now as a matter of the greatest urgency bring in a third element, the protection of human rights of all humans from Artificial Intelligence.

Necessity of Protection and Prevention

Protection of human rights of all humans has many dimensions including prevention, mitigation, and remedy of human rights violations. Given the nature of the challenges posed by Artificial Intelligence, prevention of human rights violations is of the utmost importance. The Advisory Body, and any future international agency established, must make prevention of human rights violations the first order of business. Mitigation and remedy of human rights violations may themselves have to contend with the power of artificial intelligence.

The urgency of protection of human rights, (and prevention of human rights violations in particular), makes it a matter of the greatest importance to think through the type of new international agency needed for the governance of artificial intelligence and for the protection of human rights of all humans from artificial intelligence.

It would be unwise to build a future strategy on the principle of coexistence only. A strategy of protection of humans is indispensable including the promotion of safe, secure, sustainable, and peaceful uses of these technologies and of cooperation in the field, the respect of human rights as well as the entitlement of a international agency at the UN as a regulatory authority to help determine market approval for AI products. The new international agency must be given authority to govern artificial intelligence with protection of human rights uppermost in mind.Given the similarities between Artificial Intelligence and nuclear technologies, the International Atomic Energy Agency (IAEA) would be the best model for such an institution, not least because it is one of the few UN agencies with “teeth”.

UN General Assembly Resolution: Human Rights Throughout the Life Cycle of AI

The recently adopted UN General Assembly Resolution “Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development” provides room for action on the issue of protection from AI, as is set out below.

• Room for Action by Governments as well as other Stakeholders: The General Assembly resolution on artificial intelligence is broad in scope, providing room for action by Governments as well as other stakeholders, including academic, research, and civil society organizations.
• An initial policy framework: The resolution aims for ‘safe, secure, and trustworthy artificial intelligence systems’ and sets out an initial policy framework for international cooperation on which one can build to develop an architecture of protection in this area.

• Legal and Human rights perspectives: The resolution opens by reaffirming international law, in particular the Charter of the Unted Nations, and recalls the Universal Declaration of Human Rights (first preambular para.). It emphasizes that human rights and fundamental freedoms must be respected, protected, and promoted throughout the life cycle of artificial intelligence systems (para. 5).
• Safe and secure systems: It encourages all Member States and other stakeholders to promote safe, secure, and trustworthy artificial intelligence systems (para. 6.)
• Risk monitoring and management: It specifically calls for international cooperation to foster mechanisms of risk monitoring and management. (Para. 5 (c)). It encourages research and international cooperation to understand, balance and address the potential benefits and risks related to the role of artificial intelligence systems (Para. 5(o)).
• Effective safeguards:  It calls for international cooperation for effective safeguards. (Para 5(f))
• AI governance: It acknowledges the importance of continuing the discussion on developments in the area of artificial intelligence governance so that international approaches keep pace with the evolution of artificial intelligence systems.  (Para. 8)
• Research mapping and analysis:  It “encourages continued efforts by the international community to promote inclusive research, mapping and analysis that benefits all parties…” (Para. 8).
• Development of a Global Digital Compact: It “looks forward to the development of a global digital compact” (Para. 11).
• The foregoing provisions of the GA resolution provide room for an academic/research/civil society initiative to contribute, as an initial offering, to:
  • Legal and human rights perspectives
  • Safe and secure systems
  • Risk monitoring and management
  • Effective safeguards
  • AI governance
  • Research mapping and analysis.

New International Agency Governing AI and Protecting Humans From AI

It is important to build on these provisions of the UN General Assembly Resolution in taking forward future policies and strategies in this area. As stated above, the urgency of protection of human rights, (and prevention of human rights violations in particular), makes it a matter of the greatest importance to think through the type of new international agency needed for the governance of artificial intelligence and for the protection of humans from artificial intelligence guaranteeing all humans human rights and fostering the sustainable flourishing of all humans and the planet.

Editor’s note: this post is a continuation of an earlier post, In an Era of Digital Disruptions, Ethics Can’t Be an Afterthought