Overview
- The Five Eyes intelligence alliance issued a public warning on June 22–23 that cutting‑edge AI models are likely to change offensive and defensive cyber capabilities in months and urged governments and businesses to strengthen cyber resilience now.
- The U.S. administration under President Trump ordered Anthropic to restrict non‑U.S. access to its advanced Mythos/Fable models after concerns that the models could be used to find and exploit software and system vulnerabilities.
- Major AI companies moved quickly to offer defensive products: OpenAI expanded its Daybreak security program and announced a specialized model called GPT‑5.5 Cyber designed to audit code, find vulnerabilities, propose patches, test fixes, and document corrections.
- United Nations Secretary‑General António Guterres demanded that AI firms disclose the energy, water and emissions footprints of their data centers and called for renewable power commitments by 2030 because rapid AI growth could sharply increase resource use.
- Business adoption and talent pressures are rising as firms sign major AI partnerships and surveys show many companies lack advanced operational use of AI, creating risks of 'shadow AI,' security gaps for small firms, and potential staff departures if employers fail to deliver professional AI tools.