Skip to main content

AI regulation has become urgent. IAEA Safeguards experience can provide a model for AI regulation by an International Artificial Intelligence Agency (IAIA), whose mandate would be the objective and independent verification of national AI systems. This is needed to provide assurances of the peaceful development and use of AI in participating states and to identify/prevent dangerous AI applications.

International regulation is essential to ensure the safety, reliability and transparency of Artificial Intelligence (AI) built on large language or deep learning models such as Chat GPT and GPT-4.

While the EU and China are in the process of introducing such regulation, the US has so far taken a laissez faire approach, allowing individual states and companies to decide if and how they regulate AI.

Why International AI Regulation Is Needed Now

Creating an International Artificial Intelligence Agency (IAIA) for AI regulation and for guiding AI safety and beneficial development is clearly a matter of urgency, reflected in Elon Musk’s and Warren Buffet’s prescient observations that the seriousness and magnitude of the risks associated with AI is more dangerous than those associated with nuclear weapons.

The IAIA would be responsible for international regulation of AI and the independent and objective verification of national AI to ensure compliance with national and international obligations. This verification would provide assurances that AI is being developed and used for safe and peaceful purposes only, while identifying actual/possible dangerous AI development and applications.

IAEA Safeguards as a Model for the IAIA

The experience of the International Atomic Energy Agency (IAEA) regarding IAEA Safeguards is uniquely relevant. The reasons for and role of IAIA regulation correspond essentially to those on which IAEA’s Safeguards system, international safeguards, was based. IAEA Safeguards involves the independent verification of national safeguards systems to provide credible assurances that nuclear material, items and facilities are used for peaceful purposes only.

The IAEA was established to promote the peaceful uses of nuclear energy and to verify through IAEA Safeguards that nuclear activities were not being used for non-peaceful purposes, such as the development of nuclear weapons capabilities. Similarly, the IAIA would also have the twofold task of promoting safe and peaceful uses of AI while verifying and controlling that AI is not being used for dangerous purposes.

IAIA Obligations and Benefits

  • Participating states would be responsible for declarations of all AI R&D, AI applications and other AI activities of relevant entities, all of which should adhere to the states’ legal and other commitments and obligations for the peaceful and beneficial use of AI
  • Participating states would agree to set up and train a group of qualified international inspectors empowered to inspect and verify declarations of AI activities and uncover undeclared AI activities (similar to what IAEA Safeguards inspectors do regarding nuclear materials and activities)
  • IAIA regulation would provide credible assurances that no dangerous or illicit AI aims are being pursued.
  • IAIA regulation can serve as an early warning in the event that AI is suspected of or being used for anti-social or criminal purposes
  • By agreeing to abide by international regulations and accept international verification of their AI activities, states would have access to AI technology and technical cooperation to develop their AI systems for safe and peaceful purposes
  • Serious violations of IAIA regulations could be brought to the UN Security Council (UNSC), as is the case for serious IAEA Safeguards violations

IAIA Implementation: A Roadmap

  • Participating states agree to establish IAIA charged, a., with inspection/verification of national AI systems and their compliance with relevant international norms and regulations
  • Participating states provide to IAIA declarations of their national AI activities
  • IAIA independently verifies the completeness and correctness of national AI declarations using relevant technical and other means to assure safe and peaceful AI R&D and other AI activities in participating states
  • Inspections and verifications are to be carried out throughout the various stages of AI development and application
  • IAIA would investigate and aim to uncover possible undeclared/suspicious AI activities
  • Information on AI inspections would be provided to an international AI database

International Cooperation Essential for Creating an IAIA

The main task now is whether and how to establish IAIA. It is unlikely that the AI regulations of the EU, US, China and other regions will be compatible, consistent or sufficient. There are no international regulations seriously being discussed or developed.

International cooperation was essential for creating the IAEA and international IAEA Safeguards, and it will be essential for setting up an international agency for AI regulation and for agreeing on how it would function.

A state would need to agree to relinquish a small amount of sovereignty in connection with IAIA regulation, as it does for the implementation of IAEA Safeguards. This would permit IAIA verification of national AI systems and applications as well as confirmation that a state is fulfilling its national and international AI commitments.

Setting up an international agency for AI regulation, IAIA, will require political will and the best of international diplomacy. International cooperation will be necessary to achieve international regulation for the safety, reliability, transparency of AI while ensuring and supporting the peaceful and productive activities and applications of present and future AI.

Picture © iGlobenews
WordPress Cookie Notice by Real Cookie Banner