# AI news, Wednesday, 23 September 2026

Compiled by Richard Brooks's AI news desk. Canonical: https://richard-brooks.com/ai-news/2026-09-23/

## Guterres calls for multilateral AI risk framework at UN

Policy. In his final General Assembly address he urged common safeguards and independent oversight. That adds UN-week pressure on labs and governments already debating frontier standards.

Source: Reuters, https://www.reuters.com/world/europe/un-chief-calls-ai-curbs-end-wars-his-last-assembly-address-2026-09-22/

## Salesforce ships Agent Optimizer beta for Agentforce

Agents. The tool lets teams build, test and improve production agents from session insights in one loop. Enterprise buyers get a faster path from live failures to validated changes.

Source: Salesforce, https://www.salesforce.com/blog/agent-optimizer/

## SpaceXAI releases Grok 4.7 for coding and knowledge work

Models. The new model is live in Cursor, the Grok API and GitHub Copilot at Grok 4.6 prices. It targets longer agentic coding runs and ships a stronger safeguard stack.

Source: SpaceXAI, https://x.ai/news/grok-4-7

## Twenty countries call for human control of frontier AI

Policy. Finland and Norway launched a UN-week appeal backed by 22 leaders for common standards, incident sharing and a possible global oversight body. The US and China did not sign.

Source: Office of the President of Finland, https://www.presidentti.fi/en/president-stubb-and-prime-minister-of-norway-the-rapid-development-of-ai-requires-international-cooperation/

## Xiaomi open-sources MiMo-V2.6 Pro and Flash models

Models. Pro leads open-weight agent and coding indexes at unchanged API prices. Builders get a cheaper open option for long agent runs, with RL code and task environments also released.

Source: Xiaomi MiMo, https://mimo.mi.com/docs/en-US/news/latest/v2-6