The situation
The International Telecommunication Union is convening a workshop in Chongqing, China, on 7 September alongside ITU-T Study Group 17’s 7–11 September security meetings. Hosted at the invitation of the China Automotive Engineering Research Institute, the session is intended to identify gaps and priorities for future Study Group 17 work on secure agentic-AI standardization.
The agenda covers risks including unauthorised tool or action execution, cross-agent identity spoofing, behavioural deception, goal hijacking, supply-chain poisoning and loss of autonomous control. It also addresses security-by-design, runtime controls, verification methods, and the evidence that integrators should seek from suppliers of AI agents, tools and models. Participation is open to eligible stakeholders, with remote access available after registration.
Why it matters
The meeting brings agentic-AI security into an established international telecommunications standardization process. Its immediate consequence is institutional rather than regulatory: it may shape the technical problems, assurance expectations and collaboration model considered in later ITU-T work.
No international standard, binding framework, formal work item or implementation timetable has been adopted. Any material effect on AI suppliers, deployers or cross-border governance will depend on whether Study Group 17 subsequently converts the workshop’s priorities into agreed standards work.
What changed
The ITU has moved from general discussion of agentic-AI security issues to a dedicated 7 September workshop intended to inform future Study Group 17 standardization priorities.
Impacts
Diplomacy
The workshop creates a technical forum for international and Chinese participants to shape future agentic-AI security priorities, but it has not produced a binding cross-border governance outcome.
What to watch
- Does ITU-T Study Group 17 establish a formal work item or roadmap for secure agentic-AI standardization after the Chongqing meetings?
- Do participants publish agreed security or assurance requirements for AI agents, tools and model suppliers?
Evidence and uncertainty
Low confidence. Low confidence. 1 retained references from 1 reporting origins (International Telecommunication Union). Claim-level independent corroboration has not been established. Confidence concerns the evidence for the reported development, not the probability of an outlook.
Useful context
ITU-T Study Group 17
An International Telecommunication Union telecommunication-standardization study group focused on security.
The workshop is designed to inform its future work on AI-security standardization.
Geographic context
- 01Chongqing, China (city)