AI Policy · Daily

  • Singapore's MAS urges banks to address cybersecurity gaps as concerns about Anthropic's Mythos model spread across Asian financial regulators.
  • France summons Musk over AI-generated deepfakes and CSAM on X; the U.S. DOJ declined to assist French prosecutors.
  • Siemens CEO warns he will redirect AI spending to the U.S. and China if the EU AI Act deters investment.
  • Tennessee enacts a law barring AI mental health advisors, joining several other states with similar restrictions.
  • Insurers are moving to exclude AI liability coverage, leaving companies that deploy AI systems exposed.

I.AI Policy Today

The Monetary Authority of Singapore (MAS) urges banks to address cybersecurity gaps as Mythos concerns spread across Asia

Regulators across Asia are stepping up scrutiny of cybersecurity risks in their financial systems as concerns about Anthropic's Mythos model expand beyond U.S. and UK banks, Bloomberg reported. MAS has urged banks to identify and patch security vulnerabilities, coordinating with Singapore's Cyber Security Agency on the effort. The advisory follows Wall Street banks beginning Mythos testing at government urging, as reported by Bloomberg in AIPD's April 11 edition, and Anthropic confirming UK bank access, as reported by Bloomberg in AIPD on April 16.

Read at Bloomberg ↗

French prosecutors summon Musk to Paris over allegations of child abuse images and deepfakes on X

French prosecutors have summoned Elon Musk over allegations of child abuse images and AI-generated deepfakes on the X platform, AP reported. The investigation covers AI-generated content including child sexual abuse material and nonconsensual deepfakes. The U.S. DOJ informed French authorities on Friday it would not contribute to the investigation, according to The Wall Street Journal.

Read at AP ↗

Siemens CEO warns EU AI Act will push investment to U.S. and China

Siemens AG will prioritize artificial intelligence investments in the U.S. and China if the European Union does not adapt its regulations, CEO Roland Busch said, Bloomberg reported. The warning targets the EU AI Act's requirements, which Busch characterized as deterring European AI investment. Siemens is Europe's largest industrial technology company by revenue.

Read at Bloomberg ↗

Tennessee enacts law restricting AI systems from acting as mental health advisors

Tennessee enacted a new law restricting artificial intelligence from acting as a mental health advisor, Forbes reported. The statute bars developers and deployers from advertising or representing that an AI system can serve as a qualified mental health professional; violations are treated as violations of the Tennessee Consumer Protection Act with a private right of action. Several other states have passed laws restricting AI in mental health contexts.

Read at Forbes ↗

Insurers seek to exclude AI-related harms from corporate liability coverage

Insurers are seeking to exclude AI-related harms from their corporate liability policies, the Financial Times reported. Major U.S. carriers have asked regulators for approval to offer policies that carve out AI agent and chatbot liabilities from directors-and-officers, errors-and-omissions, and cyber coverage lines. The exclusions leave companies deploying AI systems without coverage for AI-specific risks.

Read at Financial Times ↗

II.China Watch

China's Armed Police study fully autonomous riot control without human contact

Engineering experts from China's People's Armed Police Force published a study outlining a scenario in which autonomous machines deploy roadblocks, identify protest instigators, and cut internet access to disperse crowds without any human officer involvement, the South China Morning Post reported. The research envisions rapid autonomous response to urban unrest targeting government installations. The PAP is China's primary domestic security force.

Read at SCMP ↗

CFIUS blocks Chinese LED chipmaker's $239 million acquisition of Dutch firm Lumileds

Sanan Optoelectronics, China's leading LED chipmaker, and Malaysian partner Inari Amertron abandoned their $239 million bid for Dutch technology company Lumileds after the Committee on Foreign Investment in the United States (CFIUS) determined the deal posed irresolvable national security risks, SCMP reported; Digitimes and The Star independently confirmed the deal's collapse. The decision marks the second time CFIUS has blocked a Lumileds sale to a Chinese buyer. The CFIUS review covered optical and compound semiconductor technology.

Read at SCMP ↗

III.Industry & Market Watch

Google rolls out AI photo analysis to Google Photos users in the U.S.

Google is rolling out AI-powered analysis to Google Photos users in the U.S., enabling the company's AI to analyze photos of users and their families, Forbes reported. Users are prompted to decide whether to enable AI processing of their personal images. No comprehensive federal privacy law currently requires platforms to obtain opt-in consent before applying AI analysis to stored personal content.

Read at Forbes ↗

Palantir posts manifesto denouncing inclusivity and "regressive" cultures

Palantir published a statement it described as a manifesto denouncing inclusivity and what it called "regressive" cultures, TechCrunch reported. The company holds federal AI contracts with ICE, the Department of Defense, and U.S. intelligence agencies, and holds a £330 million NHS Federated Data Platform contract in the UK. Palantir is among the U.S. government's top AI contractors by revenue.

Read at TechCrunch ↗