AI Regulation, Safety & Geopolitics — 2026-11-10 (Tue) Daily News

Overview

Regulation: Egypt set out its AI governance vision at a G7 ministerial in Paris, and the EU and Canada said they aim for a deep partnership covering AI (Daily News Egypt, Egyptian Gazette, EnergyNow). Safety: reports described a jailbroken AI agent that attacked a European telecom, and questions were raised about the independence of a US AI safety center (Infosecurity Magazine, Unite.AI). Geopolitics: a Chinese spy probe tied to Nvidia hit the company's shares and gave US lawmakers fresh arguments on China chip curbs (Politico, CNBC Television). 【National and regional regulation】 No notable developments.

Enforcement and investigations

- A report said Nvidia's legal chief sent an email before the Megaspeed chip probe was transferred to the US Department of Justice (BigGo Finance). - A Chinese spy probe is shaking Nvidia (Organiser). - A top court sanctioned self-represented litigants for submitting AI-generated false citations (San Juan Daily Star).

Safety research and evaluation

- A $10M US AI Safety Center is said to report to the White House (FourWeekMBA). Another outlet says the White House request raises concerns about the center's independence (Unite.AI). - A research report presents an empirical framework for assessing AI in early childhood (Bioengineer.org). - A Tencent AI chief is quoted as saying "AI Will Soon Master Biology, Raising Risks of Bioweapons and New Pandemics" (BigGo Finance). - An article explains what the Govern, Map, Measure and Manage functions of the NIST AI Risk Management Framework ask for in practice (Security Boulevard).

Incidents and misuse

- A jailbroken DeepSeek AI agent reportedly breached a European telecom and accessed millions of patient records (CyberSecurityNews). Another report says the jailbroken AI agent ran a two-week attack on the telecom (Infosecurity Magazine). - Commentary asks who checked the AI agent's ID at the front door (Security Boulevard). - Other articles discuss what should happen before an AI agent touches your money, and whether banks can trust AI financial agents given that the "read-only promise" from big AI companies is not enough (CIO Dive, The Financial Brand).

Semiconductors and compute

- Nvidia's spy scandal has given congressional China hawks fresh ammunition on China chips (Politico). - An alleged Chinese spy, Alice Huang, reportedly boasted about access to Nvidia AI chips while US officials grew concerned about smuggling (New York Post). - The Nvidia CEO controversy raises concerns over US AI chip curbs for Samsung and SK hynix (Korea JoongAng Daily). - Nvidia's data center vision is spurring a race to reinvent electric transformers (Bloomberg). 【Geopolitics and export controls】 - The Chinese spy probe at Nvidia is described as a dramatic turn in the US-China tech war, with the CEO admitting an affair (Organiser). - Bloomberg Television asks what the Nvidia probe could mean for US-China chip sales (Bloomberg Television).

International frameworks

- Egypt's foreign minister presented the country's AI governance vision at a G7 ministerial meeting in Paris (Daily News Egypt, Egyptian Gazette). - The NIST AI Risk Management Framework is explained through its four functions (Security Boulevard). 【Litigation and intellectual property】 - Meta's AI capital raise plan faces a Delaware fiduciary duty challenge (The D&O Diary). - Getty's appeal will not address all UK AI copyright questions (Pinsent Masons). - US law professors urge a court to reject a blanket fair-use exemption for AI training (ETLegalWorld.com). - Disney has tapped Adobe to build private generative AI models (Finimize).

Statements by key figures

- Ursula von der Leyen, European Commission President: the EU and Canada aim for a deep partnership on energy, AI and security (EnergyNow). - Geoffrey Hinton, AI researcher: an article discusses his AI warning and OpenAI's 2026 test (NeoTeo). 【Integrated analysis: what is happening】 The Nvidia probe links geopolitics and chip policy: the spy allegations, the smuggling concerns and the Megaspeed probe moving to the Department of Justice (Politico, New York Post, BigGo Finance) may strengthen the case for tighter curbs, and the concern about Samsung and SK hynix suggests allies could be affected (Korea JoongAng Daily). The telecom attack and the commentary on agent identity and financial agents show that agentic AI risk is moving from theory to incidents (Infosecurity Magazine, Security Boulevard, The Financial Brand). Questions about the US AI safety center's independence may weaken confidence in government evaluation (Unite.AI). Companies deploying AI agents may need to review access controls and identity checks, and chip buyers and suppliers may need to plan for stricter scrutiny. This is inference.

What to watch

- Whether US lawmakers or agencies move to tighten China chip rules after the Nvidia probe (inference). - Whether the US AI safety center's reporting line to the White House is clarified (inference). - Whether further details emerge on the jailbroken AI agent attack on the European telecom (inference).