TFS contributed to two major United Nations initiatives focusing on the global governance of artificial intelligence.

Input to U.S. State Department Roundtable: UN Resolution on Trustworthy AI for Sustainable Development

In January 2024, TFS was invited to provide feedback on a draft UN General Assembly plenary resolution by the United States titled “Seizing the Opportunities of Safe, Secure, and Trustworthy Artificial Intelligence Systems for Sustainable Development.” The recommendations we provided in response were intended to ensure transparency and accountability in the development of advanced AI systems, enhance cybersecurity requirements for developers of advanced AI systems, and improve regulatory harmony.

Our recommendations included the incorporation of the OECD definition of AI systems, requiring developers to complete internal and third-party audits and implement Know-Your-Customer (KYC) protocols, assessing the quality of data being used to train advanced AI systems, establishing minimum diversity and labor conditions for crowd workers, and addressing a lack of diversity in AI development teams.

Following the submission of written feedback, TFS was invited by the U.S. State Department to present the recommendations at a roundtable attended by a select group of participants representing international civil society organizations, academia, and other non-governmental groups.

TFS was delighted to learn that several of our recommendations were reflected in the revised draft of the resolution in late January, including:

• Encouraging research and international cooperation to understand, balance, and address the potential risks associated with open-source AI
• Recognizing the challenge of regulation keeping pace with the rapid pace of AI development
• Facilitating comprehensive monitoring of AI systems’ impact and risks throughout their lifecycle
• Specific mention of the “prevention” of risks, rather than merely their identification, evaluation, and mitigation
• Mitigating bias encoded within datasets and combatting algorithmic discrimination
• The inclusion of information security safeguards
• Increasing inclusivity and diversity within AI R&D
• Establishing mechanisms for the reporting of AI incidents and misuse

Input to UN AI Advisory Body Interim Report: Governing AI for Humanity

TFS responded to open calls for feedback on the UN AI Advisory Body’s interim report, proposing an approach to global AI governance that addresses concerns related to the safe and ethical development and use of AI systems.

Our feedback highlighted the importance of ensuring that AI models are developed based on data sets that are relevant, representative, and, to the best extent possible, free of errors and complete, furthermore noting that insufficiently representative datasets are a fundamental cause of biased and discriminative outcomes. We also emphasized the need for robust safety measures in the development and deployment of AI systems to prevent both deliberate and accidental harm, highlighting some of the unique security risks presented by advanced open-source models. Lastly, we raised concerns about the extraordinary concentration of power within a small number of well-funded technology firms.

TFS remains committed to ensuring that AI systems are developed and used in a manner that is safe, equitable, and beneficial for all members of the international community. We are grateful for the opportunity to contribute to these important global governance initiatives and look forward to continuing our advisory role in future discussions and developments in this critical area.