Newsletter

Anthropic refuses to remove safety limits from Pentagon contract sparking a major dispute, the International AI Safety Report 2026 finds capabilities outpacing safety, Australia's data centre rush faces scrutiny, and a damning OAIC review reveals automated decision-making transparency failures.

Australia's AISI begins hiring its founding team, Grok AI deepfake crisis prompts international action, the Productivity Commission doubles down on AI regulation as a last resort, and global warnings on AI capabilities and safety gaps intensify.

Australia announces its AI Safety Institute and releases the National AI Plan, Anthropic reports AI-powered foreign espionage, and challenges to Australia's forecasted AI boom emerge around security risks and trust issues.

Australian public concern about AI grows with 65% believing it creates more problems than it solves, eSafety Commissioner targets AI chatbots with mandatory safety codes, superintelligence debate goes mainstream, and Treasury takes a light-touch approach to AI regulation.

Australia faces a crossroads on AI regulation as debate over a national AI act intensifies, Trump releases comprehensive AI Action Plan, China unveils competing governance vision, GPT-5 highlights frontier risks, and UK AISI conducts landmark pre-release testing.

EU publishes GPAI Code of Practice, Switzerland develops sovereign multilingual LLM, Trump releases AI Action Plan, and OpenAI publishes economic blueprint for Australia.

Newsom's AI working group releases policy guidelines report, G7 leaders issue statement on AI for prosperity, Albanese signs Amazon data centre deal, and tech giants shift focus to superintelligence.

Trump dismantles Biden administration AI safety initiatives, OpenAI announces the Stargate project expansion to US allies, and the Federal Court of Australia begins consultations on generative AI use in legal proceedings.

Australian Senate inquiry releases its AI report with recommendations for dedicated legislation, Singapore bans AI-manipulated election content, and Canada announces its national AI Safety Institute.

Australia progresses voluntary AI safety standards and mandatory guardrails, Newsom vetoes SB 1047 but signs other AI safety bills, and the UN releases practical proposals for AI governance.

Australia bans deepfake sexual material, the US Senate considers new AI content protection bills and regulatory sandboxes for financial firms, and California's SB 1047 advances.

California advances regulation for costly AI models, Australian governments release a joint AI framework, and creating sexual deepfake images without consent is criminalised in Australia.