š¤ AI Policy this week #046. Debates at the EU Commission to make some rules voluntary; AI security council launched in Korea.
A quick summary of news, reports and events discussing the present and future of AI and the governance framework around its development.
Returning from some hectic weeks (stay tuned for what is comingā¦). The new wave to soften AI rules keeps advancing worldwide: the EU Commission plans to make parts of the AI code of practice voluntary, the US studies its Trump2.0 Action Plan while the Virginia Governor vetoes an AI Bill, and Japan plans to adopt āAI-Friendly Legislationā. All this, surrounded by Studio Ghibli AI-generated images sparking comments on copyright and privacy.
1. News
EU Commission plans to change AI code of practice spark debate.
Seven MEPs have written to the European Commission in Brussels saying they are āgreatly concernedā about plans to change the code of practice enshrined in new legislation to govern the use of AI. They have been working on the legislation together with the commission and member states. But now, they say, the commission has proposed watering down the code to make it entirely voluntary for the providers of general-purpose AI systems when it comes to assessing and mitigating risks to fundamental rights and democracy.
German coalition disagrees on AI regulation.
Leaked coalition documents reveal disagreement between the centre-right Christian Democrats (CDU/CSU) and the centre-left Social Democratic Party (SPD) working groups on AI regulation and digital sovereignty ambitions in negotiations over a new German government platform. The leaked draft, with statements marked as CDU/CSU, SPD or unity, favours a digital sovereignty push and "innovation-friendly" implementation of AI regulation, but splits on the ambitions of the twin agendas.
US Senate confirms Michael Kratsios to lead White House science, technology office.
Senators voted 74-25 to approve Michael Kratsios, who served as Trumpās chief technology officer in his first administration. Kratsios, who will serve as both OSTP director and assistant to the president for science and technology, previously served as the managing director of Scale AI. Kratsios will take the lead at OSTP while the office is involved in the development of the administrationās AI policy.
Virginia Governor Vetoes Artificial Intelligence Bill HB 2094.
Virginia Governor Glenn Youngkin vetoedAI-focused House Bill 2094, which proposed requirements for the development, deployment and use of high-risk AI systems and laid out penalties for noncompliance. The state-level legislative move is the latest sign of a push toward minimizing government oversight as the technology develops. Youngkin characterized the bill as aburdensome regulatory framework that would undermine progress and stifle economic growth.
OpenAI peels back ChatGPTās safeguards around image creation.
OpenAIlaunched a new image generator in ChatGPT, which quickly went viral for its ability to createStudio Ghibli-style images (Sam Altman says ChatGPTās viral image-generation AI is āmeltingā OpenAIās GPUs). Beyond the pastel illustrations, GPT-4oās native image generator significantly upgrades ChatGPTās capabilities, improving picture editing, text rendering, and spatial representation. However, one of the most notable changes OpenAI made this week involves its content moderation policies, which now allow ChatGPT to, upon request, generate images depicting public figures, hateful symbols, and racial features.
ASCAP lobbies for āHumans Firstā approach in AI copyright policy.
The American Society of Composers, Authors, and Publishers (ASCAP) is calling for a āHumans Firstā approach in the somewhat concerningly evolution of AI and copyright law. In response to the US administrationās call for feedback on its āAI Action Planā, ASCAP has voiced concerns about the potential implications for human creators.
Singaporeās senior civil servant overseeing AI and Smart Nation resigns ahead of the 2025 general election.
Ms Goh Hanyan, a director in the Smart Nation Strategy Office and the national AI group for policy and strategy in the Ministry of Digital Development and Information (MDDI), will leave the service on April 3, The Straits Times has learnt. She is the third high-ranking civil servant known to step down ahead of the election so far, following the resignations ofdeputy secretary Jasmin Lau from the Ministry of Health andMr Foo Cexiang, a director from the Ministry of Transport. Ms Goh was involved in shaping Singaporeās refreshed Smart Nation plan, known as Smart Nation 2.0, as well as the coordination and implementation of the Republicās refreshed National Artificial Intelligence Strategy.
Korea launches national AI security council to address cybersecurity threats.
The presidential office's National Security Office announced on the 28th that it has launched the 'National AI Security Council (NACG)' to establish national-level responses to security threats related to artificial intelligence (AI) and held a joint meeting of relevant agencies. Officials at the director general level from seven government agencies, including the National Intelligence Service, Ministry of Science and ICT, Ministry of Foreign Affairs, Ministry of National Defense, Ministry of the Interior and Safety, Personal Information Protection Commission, and Digital Platform Government Committee attended the meeting.
South Korean Trade Group Forms AI Policy Committee.
A trade group representing major South Korean companies, including Samsung, Hyundai Motor, SK, and LG, has established an artificial intelligence (AI) policy committee. The committee will likely focus on developing strategies and recommendations related to AI ethics, governance, and innovation. By creating this committee, the trade group aims to foster collaboration among its members, share best practices, and develop common positions on AI-related issues.
Japan Plans to Adopt AI-Friendly Legislation; AI draft bill is already in Congress.
The Japanese Governmentannounced its intention to position Japan as āthe most AI-friendly country in the worldā, with a lighter regulatory approach than that of the EU and some other nations. This statement follows: (i) the Japanese governmentās recent submission of an AI bill to Japanās Parliament, and (ii) the Japanese Personal Data Protection Commissionās (āPPCā) proposals to amend the Japanese Act on the Protection of Personal Information (āAPPIā) to facilitate the use of personal data for the development of AI. On February 28, 2025, the Japanese Governmentsubmitted its āBill on the Promotion of Research, Development and Utilization of Artificial Intelligence-Related Technologiesā (āAI Billā) to Parliament. If enacted, it would become Japanās first comprehensive law on AI.
Amazon Changes Alexa Policy To Enhance AI Features
Amazon implemented a significant policy change on March 28, 2025, that will affect how Echo devices manage user privacy. All Alexa conversations will be sent to the cloud, a shift aimed at enhancing the capabilities of Amazon's new generative AI (genAI) powered assistant, Alexa+. This decision comes as part of a broader strategy to revitalize Alexa, which has already been integrated into over 600 million devices worldwide but has faced challenges in monetization. The policy change will eliminate the existing 'Do Not Send Voice Recordings' option, which allowed users to keep their voice commands local to the device. Instead, users will be given a new option that permits Amazon to delete recordings from the cloud at an unspecified later date.
2. Reports, Briefs and Opinion Pieces:
āDraft Report of the Joint California Policy Working Group on AI Frontier Modelsā.
The report āleverages broad evidence to provide a framework for policymaking on frontier artificial intelligence (AI), taking into account the importance of innovation and establishing appropriate strategies to reduce material risks. It surveys AI research representing multiple disciplines to derive policy principles that can inform how California approaches the use, assessment, and governance of frontier AIā.
āThe Forthcoming Artificial Intelligence Action Planā by the American Action Forum.
Angela Luna, Tech & Innovation policy analyst, summarizes the responses submitted to the US Plan. āFive core themes emerged as significantly overlapping across organizations: export controls and U.S. leadership infrastructure investment, regulatory frameworks, ethical artificial intelligence (AI) development, safety and security, and intellectual propertyā.
3. Events:
SEC Roundtable on Artificial Intelligence in the Financial Industry (Mar 27, Washington DC, US).
Acting Securities and Exchange Commission Chairman Mark Uyeda called for a softer agency approach to artificial intelligence regulation, reflecting a broader push in the Trump administration. Speaking at the AI event hosted by the SEC in Washington, Uyeda said the agency should generally avoid āoverly prescriptiveā AI rules that can quickly become outdated and impede innovation.
āArtificial Intelligence Landscape Assessment (AILA) in the Public Sector in Viet Namā (Mar 18, Hanoi, Viet Nam).
The thematic seminar, co-organized by the United Nations Development Programme (UNDP) and the Institute for Policy Studies and Media Development (IPS), provided āan in-depth assessment of the current AI landscape, strategic directions, and application status in the public sector, while offering policy recommendations to fully leverage AI's potentialā.
Thanks for reading, please share any comments and see you next week.