The International Telecommunication Union, the United Nations agency for digital technologies, has announced a new initiative to develop frameworks for trusted digital identity and to ensure the behaviour of AI agents remains trustworthy and accountable throughout their lifecycle. The announcement came at the AI for Good Global Summit in Geneva, part of a wider Digital Week of UN-linked technology events running through July 10.
AI agents represent a new generation of systems designed to act independently on behalf of users, carrying out tasks ranging from scheduling and purchasing to complex business processes. While they promise productivity gains, the ITU flagged the risk of such agents impersonating people or taking unauthorised decisions.
A Focus Group for an Emerging Class of Digital Actor
To address these risks, the ITU will establish a Focus Group on Trust and Identity for Humans and Agentic AI, tasked with developing frameworks that preserve meaningful human control over tasks such as executing financial transactions and operating critical infrastructure. The group will comprise technical, policy and legal experts, holding its first meeting in Paris in November and its second in Geneva in January.
Focus Group Co-Chair Debora Comparin framed the urgency plainly. She said AI agents will soon negotiate, transact and make decisions on people’s behalf, and that common international foundations are needed to establish who the agents are and how and when they can be trusted. Co-Chair Amir Banifatemi added that identity establishes who is acting while trustworthiness determines how that actor can be expected to behave, and that bringing the two together creates the foundation needed for interoperable, accountable AI systems at global scale.
The Focus Group will report to the ITU’s expert group for security standards, known as ITU-T Study Group 17. Study Group 17 Chair Arnaud Taddei said the effort was moving strategically and swiftly while deliberately taking time to get the foundations right, adding that the direction was clear even as the leadership and structure were still being assembled.
Part of a Larger Geneva Push on AI Governance
The ITU initiative did not emerge in isolation. Days earlier, the same summit saw the launch of the AI for Good Global Commission, co-chaired by Salesforce’s Marc Benioff and Rwandan President Paul Kagame, bringing together more than 40 founding members including heads of state and government, industry CEOs and heads of UN agencies to define practical pathways to strengthen trust, expand access and unlock AI’s potential. The commission cited the statistic that 2.2 billion people remain offline, cut off entirely from AI’s advances, as central to its rationale.
Together, the two initiatives illustrate a shift in how multilateral bodies are approaching AI: not as a single governance problem to be solved once, but as a layered set of technical, economic and trust-related challenges requiring separate specialised tracks working in parallel.
Why India Has a Direct Stake
The concerns driving the ITU’s Focus Group are not abstract for Indian regulators. The Reserve Bank of India released its own draft Guidance on Regulatory Principles for Model Risk Management on June 24, requiring every bank and NBFC to build “kill switch” mechanisms capable of instantly overriding or deactivating any AI model, alongside mandatory human oversight for high-risk automated decisions.
That draft, open for public comment until July 24, was triggered partly by concerns over increasingly autonomous, agentic AI systems being used to probe banking infrastructure. With UPI having recorded 106 billion transactions in just the first half of 2025, and AI-driven fraud detection now embedded across nearly every major Indian bank and NBFC, the stakes attached to establishing which AI agents can be trusted, and which cannot, are considerable.
The ITU’s identity-and-trust framework, once developed, could eventually offer Indian regulators a common international reference point, aligning domestic rules like the RBI’s kill-switch mandate with an emerging global standard for how autonomous agents identify themselves and remain accountable, rather than leaving India to build such standards in isolation.
