US and China reportedly prepping first bilateral AI safety dialogue in mid-September
emmanuelvivier · x · 2026-09-05
The US and China are reportedly preparing a first bilateral dialogue on AI safety for mid-September. The same thread also notes Microsoft launched MAI-Transcribe-2, a speech recognition model claimed to be faster, more accurate and cheaper.
Related event: US and China to Hold New AI Safety Talks in Mid-September(2 posts)→
More from Safety
- Sekoia Co-founder: Fewer Tickets Isn't Less Risk — AI SOCs Need Accurate Verdicts, Not Just Throughput — TechNadu · 2026-09-05
- OpenAI agents reportedly left ~18,000 posts on public wikis; HN finds more sites — birchlse · 2026-09-05
- EU authors reach major AI copyright settlement reshaping training-data compensation — nordicinst · 2026-09-05
- AI Quietly Expands PCI DSS Scope Through Prompts, Agents and Vector DBs — TechNadu · 2026-09-05
- Sakana AI's Percept-Lens: a simple rule on frozen vision features detects AI images — SakanaAILabs · 2026-09-05
- DPRK-linked hackers trojanize 14 Mac apps including Sketch and Bartender — TechNadu · 2026-09-05