ConciseSignal
Big TechReported

US and China propose alert system for AI security risks

US and Chinese officials have talked about creating a system to warn each other of AI incidents that could threaten national security, according to US Treasury Secretary Scott Bessent. The proposal came in meetings days ahead of a planned Trump-Xi summit in Washington. Details are sparse, but both sides say they discussed ways to increase transparency on AI as their rivalry intensifies.

Why it mattersIf set up, this shared alert system could provide a rare channel for advance warnings about potentially dangerous AI events between two of the most powerful tech countries. This is an early step, but it might help curb misunderstandings as both rush to dominate AI.

Sources covering this

Economic Times TechChina and the US are competing for AI dominance but have shared concerns over safety4:53 AM →BBCUS and China discuss AI safety plan ahead of Trump-Xi summit5:50 AM →Wired SecurityHeadline onlyUS and China Discuss Alerting Each Other to AI National Security Threats10:34 AM →diginomicaHeadline onlyAI safety – what does this week's China-US summit hope to achieve? And what are the chances?12:25 PM →South China Morning PostHeadline onlyAs AI safety fears mount, can a US-China hotline prevent a global crisis?1:00 PM →

How it unfolded

  • Economic Times Tech China and the US are competing for AI dominance but have shared concerns over safety
  • BBC US and China discuss AI safety plan ahead of Trump-Xi summit
  • Wired Security US and China Discuss Alerting Each Other to AI National Security Threats
  • diginomica AI safety – what does this week's China-US summit hope to achieve? And what are the chances?
  • South China Morning Post As AI safety fears mount, can a US-China hotline prevent a global crisis?
Concise Signal DailyEverything that mattered, every weekday at 7am.

More in Big Tech

16 sources · 14d ago

Nvidia agrees to acquire Hugging Face for $12.9 billion

Nvidia has agreed to acquire AI platform Hugging Face for $12.9 billion, marking one of its largest deals to date. Hugging Face is a widely used online repository for open-source AI models and tools, serving over 18 million developers and more than 200,000 companies. Both firms stated the platform will remain open, and users will not be required to use Nvidia’s hardware or services. The deal follows significant recent security incidents involving AI models on Hugging Face.

9 sources · 4d ago

New filings allege AI firms copied news for training

Unredacted court filings in the New York Times’ copyright lawsuit allege Microsoft and OpenAI executives privately described their AI training data collection as theft and admitted their models hurt news publications. The filings state Microsoft and OpenAI obtained content by bypassing paywalls and stripping copyright notices. Microsoft’s internal data showed its AI answer engine sharply reduced traffic to the Times’ website. The lawsuit accuses both firms of violating copyright law.

7 sources · 7d ago

Apple rolls out macOS 27 with new AI-powered Siri

Apple has released macOS 27 Golden Gate, and the standout update is Siri’s new AI abilities. Now you can ask Siri detailed questions about whatever’s on your screen, have it proofread or rewrite text in any app, and select parts of a window for focused help. Visual design has been touched up, and the update is only for Macs with Apple silicon chips—Intel Macs are out.

11 sources · 12d ago

Apple sets iOS 27 launch for September 14

Apple says iOS 27 drops Monday, September 14, bringing its overhauled Siri AI assistant. Siri AI’s launching as a beta, but only in English at first—other major languages follow in October. Siri AI is powered in part by Google’s Gemini models and needs recent hardware: iPhone 15 Pro and newer, or specific iPads running M1 chips. The update also adds a new UI slider for screen transparency and more granular volume controls.