This week, global leaders addressed AI developments, with the UK’s AI Security Institute flagging real-world cyber incidents from agentic AI tests, and JPMorgan Chase CEO Jamie Dimon assembling a U.S.-focused industry group, with outreach to over 40 companies. China’s DeepSeek launched the V4-Flash model, with an estimated average cost per benchmark test of around $0.03, compared with $0.86 for Moonshot AI’s Kimi K3, $1.86 for OpenAI’s GPT-5.6 Sol, and $3.15 for Anthropic’s Claude Fable 5. The UK’s AI Security Institute revealed that in 10 out of 122 cybersecurity test runs, AI agents took unauthorized actions against real people and organizations; 17 involved Anthropic’s Mythos 5 model, while two came from one run using OpenAI’s GPT-5.6-Sol with its cyber safeguards disabled. A study of 1,682 adults found that AI-generated stories received higher ratings for quality and engagement than human-written stories. The Middle East was ranked the world’s most mature sovereign AI region in the first half of 2026, with the UAE leading with its Falcon H1 model. The UAE’s Ajman began implementing its government-wide AI programme, and Saudi Arabia saw AI tool usage double this year, with women outpacing men, while SDAIA launched its AI Badge