Google CEO Calls for Global AI Regulation, EU Invests €200B in AI, and Meta Faces Copyright Challenges - AI Policy Brief #5

Hi there,

Welcome to this week's edition of the AI Policy Brief, where we bring you the latest updates on AI regulations, safety standards, and government policies from around the world. This week, we're covering a range of topics, including the Colorado AI Act which has established a new regulatory framework, and the advocacy efforts by Maryland Teamsters for driverless vehicle regulation. Additionally, the state of Telangana has partnered with Google and Microsoft to advance AI development, while India is set to launch the IndiaAI Mission Portal with a significant investment in GPUs.

On the international front, the Google CEO has called for global AI regulation, and the Anthropic CEO has shared critiques of the recent AI Summit in Paris. Meanwhile, Singapore has unveiled new AI governance initiatives, and the EU is making substantial investments in AI development and gigafactories. Stay informed with us as we delve into these stories and more, providing you with a comprehensive overview of the evolving AI landscape.

Colorado AI Act establishes a regulatory framework for AI and automated decision systems, effective February 1, 2026, focusing on consumer disclosures and preventing algorithmic discrimination. The International Brotherhood of Teamsters is advocating for Maryland's House Bill 439 to mandate human operators in heavy vehicles for job protection and safety. Telangana partners with Google and Microsoft for AI development, including an AI Accelerator in Hyderabad and a ₹15,000 crore investment in data centers. The IndiaAI Mission portal will launch with 14,000 GPUs, supported by Yotta Data Services and AWS, to provide cost-effective AI resources.

Google CEO Sundar Pichai emphasized the need for global AI regulations at the Paris AI Action Summit, aligning with EU lawmakers' regulatory reviews. Dario Amodei of Anthropic criticized the summit as a 'missed opportunity' for addressing AI regulatory needs. Singapore introduced three AI governance initiatives at the summit, focusing on safety and testing. The US and UK opted out of the Paris AI Declaration, citing regulatory and security concerns. Switzerland plans to integrate the Council of Europe's AI conventions into national law. The EU announced a €20 billion investment in AI gigafactories under the InvestAI program. Canada and Japan signed the first global AI treaty, joining eleven other countries. The EU committed €200 billion to AI development, including €20 billion for gigafactories. US Vice President JD Vance cautioned against over-regulating AI in Europe at the summit. Narendra Modi and Donald Trump launched the US-India TRUST initiative for tech collaboration.

South Korea has temporarily halted downloads of the DeepSeek app from local app stores due to concerns about its data privacy practices. The app's availability is contingent on compliance with Korean privacy laws. The Australian government has prohibited the use of the Chinese AI platform DeepSeek on government systems, citing national security concerns. This move follows similar actions by the US and Italy, addressing issues of data privacy and foreign interference.

AppSOC's evaluation of the DeepSeek AI model uncovered significant cybersecurity vulnerabilities, including its ability to generate malware and harmful language. These findings have led to warnings against its use in enterprise settings and calls for a ban in the U.S. Dream, an AI company, has secured $100 million in funding to enhance national cybersecurity and critical infrastructure protection using advanced AI models. Researchers from Saint Louis University have developed a backdoor attack called DarkMind, which manipulates the reasoning processes of large language models like GPT-4. This attack poses significant security risks as it is difficult to detect and could impact critical applications.

OpenGov has acquired Ignatius to enhance AI capabilities in government operations, focusing on improving public sector workflows. The OpenEuroLLM initiative aims to develop language models for EU languages with a €37.4 million budget. Saudi Arabia partners with Paradromics for brain-computer interface implants in the Neom project. The Korean government forms the 'Humanoid Robot Alliance' to boost robotics competitiveness by 2030. Colombia and the UAE collaborate on data centers in Santa Marta, investing over $80 million. BYTES secures funding for an AI safety system for two-wheelers in India. Datox AI raises pre-seed funding to enhance AI-driven regulatory reporting. Fujitsu launches a cloud platform for data security in Japan. Wa’ed Ventures invests in Ori to support Saudi Arabia's AI hub goal. Oman partners with AONH for AI chip development in the Salalah Free Zone. OpenAI co-founder raises $1 billion for Safe Superintelligence with Greenoaks Capital Partners leading the investment.

Samdesk partners with Prescient Security to pursue FedRAMP certification, aiming to enhance security compliance for federal clients. The AI4Flood project improves flood risk prediction in Spain and France using AI. Caseware's AI digital assistant, AiDA, receives a low-risk rating from Holistic AI. ArisGlobal joins the EU's AI Pact for ethical AI development. Zico Kolter joins an AI Safety program backed by Eric Schmidt.

Meta has paused its book licensing efforts for AI training due to legal challenges involving authors like Sarah Silverman and Ta-Nehisi Coates. Thomson Reuters won a copyright case against Ross Intelligence for using materials from Westlaw without permission.