top of page
actual logo.png

Regulatory Updates Newsletter: February 2025

Updated: Mar 6

February was a month dominated by AI governance as various regulatory bodies scrambled to put in place some ground-rules, in the face of a slew of headline grabbing LLM model releases. Notably, Japan came out with a cautious stance on AI risks, just as the first rules under the EU AI Act regulations started to apply.  On the other hand, the AI action summit at Paris provided the first glimpses of the divergence in views surrounding AI regulations that could get more pronounced in the future.

 

In other developments, the SEC has started signaling a major revision in crypto enforcement actions, while enhancing focus on cybersecurity and fraud in the digital assets space. The Fed kickstarted its annual CCAR exercise by releasing hypothetical scenarios, while OSFI (Canada) decided to defer Basel III capital buffer applicability to 2026, and the EU proposed shorter settlement periods for securities transactions.

 

Dive into some of these important updates below.

In the end, we also enclose a summary table for additional updates with links to all original source material.



EU initiatives on AI and other Regulations


  1. EU publishes guidelines on prohibited artificial intelligence practices

Certain provisions under the EU AI Act have started to apply, including the AI system definition, certain AI literacy requirements, and a very limited number of prohibited AI use cases that pose unacceptable risks in the EU.  


Regulators are taking concrete steps to ensure adherence by releasing guidelines surrounding prohibited AI practices and unacceptable risks. The guidelines provide legal explanations and practical examples to help stakeholders understand and comply with the AI Act's requirements.

Source: https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-ai-system-definition-facilitate-first-ai-acts-rules-application

  1. EU Commission publishes Work Programme 2025

The European Commission’s 2025 Work Programme prioritizes advancing AI governance and digital infrastructure including focus on a Single Market Strategy, a revision of the Sustainable Finance Disclosure Regulation, Digital Networks Act and investment simplification.

Source: https://commission.europa.eu/document/download/f80922dd-932d-4c4a-a18c-d800837fbb23_en?filename=COM_2025_45_1_EN.pdf

  1. EU Transactions in Transferrable Securities

The EU Commission has proposed amending the Central Securities Depositories Regulation (CSDR) to shorten the settlement period for securities transactions from T+2 to T+1, effective October 11, 2027. The transition timeline aims to allow time for testing, and coordination ensuring a smooth implementation.


Source: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52025PC0038

Paris AI Action Summit Update

SEC’s February 2025 Crypto Shake-Up

Canada’s OSFI’s Final Quarterly Release Pilot

RBI releases regulatory handbook- “Regulations at  a Glance”

Federal Reserve Board releases the hypothetical scenarios for its annual stress test


Japan’s AI Regulation: Balancing Innovation and Risk

On February 4, 2025, the Japanese government’s Cabinet Office released the interim report of its AI Policy Study Group.


  1. Businesses, Startups and AI Risks

Unlike earlier proposals for sweeping AI regulations, Japan’s Interim Report (Feb 2025) opts for a cautious, sector-specific approach by using existing laws and voluntary risk mitigation by businesses, while the government continuously monitors emerging risks.  It recommends forming a strategic leadership body to coordinate policy and incident information—without imposing legal sanctions—highlighting the complex nature of AI risks.

 

  1. Japan’s response on DeepSeek Shock

On February 6, 2025, the Japanese government issued an advisory regarding the use of DeepSeek. This notice primarily highlighted that data acquired by DeepSeek is stored on servers in China. Hence, subject to Chinese legal jurisdiction.

 

  1. Business-led AI Governance

The Interim Report reveals that Japan plans to lean on voluntary business measures to manage AI risks under current laws. However, this approach may not be sufficient for all companies—especially startups, which often lack the resources needed for robust safety and governance.

 

Source: https://www.csis.org/analysis/new-government-policy-shows-japan-favors-light-touch-ai-regulation

Summary table of additional updates



For feedback or regional coverage requests, contact us at info@riskinfo.ai.

 


Best regards, 

The RiskInfo.ai Team


Comments

bottom of page