Securiti leads GigaOm's DSPM Vendor Evaluation with top ratings across technical capabilities & business value.
View
August 26, 2025
Colorado, United States
The Colorado legislature has voted to extend the effective date of the Colorado AI Act- the first comprehensive state AI law in the U.S. from February 1, 2026, to June 30, 2026. The move came after lawmakers were unable to reach a consensus on proposed amendments during a special session held from August 21-26.
Despite four amendment bills being introduced, disagreements among legislators, civil society, and a divided technology and business lobby prevented compromise. The delay provides additional time for debate when the legislature reconvenes on January 14, 2026, ahead of the law’s new implementation deadline.
August 1, 2025
Illinois, United States
Illinois has enacted the Wellness and Oversight for Psychological Resources Act, one of the first state laws to regulate AI in therapy. The Act, effective immediately, prohibits AI systems and chatbots from delivering professional therapy or making therapeutic decisions. Only licensed professionals may provide such services, and they may not delegate independent therapeutic communication or recommendations to AI.
AI can still be used for administrative tasks, such as scheduling and billing, and for limited supplementary support like clinical documentation, but only with patient consent. Violations carry penalties of up to $10,000 per incident. The law reflects growing concern over AI-powered chatbots providing harmful or inaccurate mental health advice.
August 27, 2025
The European Commission has sought feedback on how the Digital Markets Act (DMA) ensures free competition within the digital markets, including the AI sector, in its call for evidence.
This review is aimed at assessing the law’s effectiveness as well as potential areas for improvement. Any feedback gained will be used to create a report to be presented to the European Parliament, Council, and the European Economic and Social Committee in May 2026. The deadline is set to September 24, 2025, for all feedback.
August 11, 2025
Germany & France
The Federal Office for Information Security (BSI) in Germany and the Cybersecurity Agency (ANSSI) in France have released a joint statement on securing AI systems through zero trust principles. Their paper, “Design Principles for LLM-based Systems with Zero Trust,” addresses challenges posed by LLMs by extending the traditional zero trust approaches to AI-specific threats.
The guidance recommends protection of sensitive AI components like model weights and training data from any unauthorized access, continuously monitoring AI inputs and outputs for suspicious activity, implementing defenses against AI attacks like data poisoning and model evasion, limiting AI system access rights, ensuring transparent AI decision-making, and maintaining human oversight for critical decisions.
August 6, 2025
The European Insurance and Occupational Pensions Authority (EIOPA) has published an Opinion clarifying how existing insurance-sector legislation applies to the use of AI systems. The guidance aims to help national supervisors and insurers interpret frameworks such as the Insurance Distribution Directive and Solvency II Directive in light of the EU AI Act, which classifies AI systems used in life and health insurance pricing and risk assessment as high-risk.
The Opinion does not create new obligations but sets supervisory expectations around governance and risk management. It emphasizes principles such as data governance, record-keeping, fairness, cybersecurity, explainability, and human oversight, following a risk-based and proportionate approach. EIOPA also intends to issue further guidance on specific AI use cases and emerging issues in the sector.
August 4, 2025
The European Commission has released the list of 25 organisations that have signed the General-Purpose AI (GPAI) Code of Practice, a voluntary framework supporting compliance with the AI Act. Signatories include global leaders such as Google, Microsoft, OpenAI, Anthropic, Mistral AI, Amazon, IBM, and Cohere, alongside several European AI firms.
Notably, X signed only the Safety and Security chapter, meaning it must demonstrate compliance with transparency and copyright obligations through alternative methods. The Commission’s publication of the list underscores growing industry alignment with the EU’s approach to AI accountability and governance.
August 2, 2025
The AI Act obligations for providers of general-purpose AI (GPAI) models have officially taken effect across the EU, ushering in new requirements for transparency, copyright, and responsible development. Providers must now disclose training data summaries, safeguard copyright, and ensure models meet baseline safety standards.
The Commission has issued guidelines clarifying compliance duties and confirmed the GPAI Code of Practice as a voluntary tool to ease implementation and provide legal certainty. While most providers must meet transparency and copyright obligations by August 2, 2025, models already on the market have until August 2, 2027, to comply. More advanced models with systemic risks face stricter obligations, including Commission notification and enhanced safety measures.
August 1, 2025
The EU’s AI Board has formally approved the General-Purpose AI (GPAI) Code of Practice, confirming it as an adequate voluntary mechanism for providers of GPAI models to demonstrate compliance with the AI Act. Published in July 2025 after a multi-stakeholder drafting process, the Code addresses obligations around transparency, copyright, and safety.
By adhering to the Code, GPAI providers can reduce administrative burdens and gain greater legal certainty when placing their models on the EU market. The approval highlights the EU’s push to create practical pathways for compliance as GPAI obligations under the AI Act begin to take effect.
August 1, 2025
The European Parliamentary Research Service (EPRS) has released a report examining the copyright implications of generative AI in the context of large language model (LLM) training. The study highlights two central challenges: attribution, i.e., whether outputs meaningfully derive from training data, and novelty, i.e., whether outputs represent genuine new creations or statistical reproductions of existing works.
The report recommends three key steps: requiring AI developers to track and disclose training data with independent audits, creating compensation systems that reflect the statistical influence of creators’ work, and developing open standards to ensure collaboration between creators, regulators, and researchers. EPRS concludes that the EU is well-positioned to set global standards for transparency, attribution, and accountability in generative AI development.
August 22, 2025
China
China’s Ministry of Industry and Information Technology (MIIT) has issued draft Administrative Measures for the Ethical Management of AI for public comment. The rules apply to AI R&D and applications that pose risks to life, health, dignity, the environment, or public order.
They establish a four-tier review system including expedited reviews within 72 hours for urgent cases and require institutions to register projects, report to a national platform, and comply with oversight or face penalties. The draft highlights China’s push to pair rapid AI innovation with ethical governance and global standard-setting.
August 13, 2025
India
The Reserve Bank of India (RBI) has published the report of its Committee on the Framework for Responsible and Ethical Enablement of Artificial Intelligence (FREE-AI) in the Financial Sector. Established in December 2024, the Committee engaged with a broad set of stakeholders before finalizing its recommendations.
The report sets out 7 “Sutras” as guiding principles for AI adoption in finance, supported by 26 actionable recommendations across six strategic pillars. The framework emphasizes balancing innovation with risk management, ensuring AI deployment enhances efficiency, trust, and resilience without compromising consumer protection or financial stability.
The report is now available on the RBI’s website for wider review and discussion.
August 11, 2025
Nepal
Nepal has adopted its first National AI Policy, 2025, aimed at fostering ethical, secure, and inclusive use of AI while supporting research, startups, and public–private partnerships. The policy establishes legal and regulatory frameworks, calls for an AI Regulation Council and a National AI Centre, and commits to reviews every two years.
By expanding infrastructure such as data centres and high-speed connectivity, and embedding safeguards for privacy, rights, and security, the policy seeks to integrate AI across key sectors and position Nepal as a competitive player in global AI markets.
August 11, 2025
Indonesia
Indonesia’s Ministry of Communications & Informatics has launched a public consultation on its National AI Roadmap and Draft AI Ethics Guidelines. Drawing from global standards set by UNESCO, OECD, and ASEAN, the guidelines emphasize inclusivity, transparency, and accountability in AI development and use.
By embedding these international principles, Indonesia aims to position itself as a regional leader in shaping responsible AI governance across ASEAN. Public feedback is open untilAugust 22, 2025.
August 8, 2025
Saudi Arabia
The Saudi Data & AI Authority (SDAIA) has published a report on Agentic AI, defining its six core capabilities: perception, reasoning, learning, action-taking, communication, and autonomous operation. The report charts AI’s evolution from early rule-based systems to today’s generative agents powered by large language models, which can collaborate to achieve complex goals.
Framing Agentic AI as a pillar of Saudi Vision 2030, SDAIA highlights its potential to accelerate digital transformation and support the Kingdom’s transition to a knowledge-driven economy.
August 4, 2025
At the first-ever APEC Digital and AI Ministerial Meeting in Incheon, ministers from across APEC economies issued a joint statement committing to responsible digital transformation and trusted AI development. The statement emphasized three priorities: advancing AI innovation to address socio-economic challenges, expanding digital connectivity, and ensuring a safe and reliable digital ecosystem.
Ministers also recognized Korea’s leadership in spearheading a new APEC AI initiative, expected by the end of 2025, and pledged to sustain momentum under the APEC Internet and Digital Economy Roadmap.
August 1, 2025
Australia
The Office of the Australian Information Commissioner (OAIC) has closed preliminary inquiries into I-MED Radiology Network, Harrison.ai, and Annalise.ai regarding the disclosure of medical imaging data for AI model training.
The OAIC reviewed whether I-MED’s sharing of patient data between 2020 and 2022 complied with the Australian Privacy Principles (APPs). The Commissioner found that the data had been sufficiently de-identified and was therefore no longer considered personal information under the Privacy Act.
While no regulatory action will be taken, the OAIC emphasized that developing AI models with large datasets remains a high-risk activity and highlighted this case as an example of good de-identification and contractual safeguards.
Taiwan’s Generative AI Competition Review: The Fair Trade Commission’s consultation on generative AI’s impact on competition, covering hardware supply chains, deployment, and risks like dominance and collusion, is open until September 7, 2025.
China’s AI Content Labeling Rules: Effective September 1, 2025, platforms must embed visible labels and metadata on AI-generated text, images, audio, and video, while also flagging suspected synthetic content.
Croatian and German data protection authorities’ (AZOP-BfD)I Guidelines on Personal Data Protection in AI Trainings: Croatian and German data protection authorities are consulting until August 30, 2025, on protecting personal data in AI training, particularly around memorization risks in large language models. The results may influence Europe-wide standards.
California’s Advances 3 Major AI Bills: California is advancing three major AI bills: Assembly Bill 1064 (Leading Ethical AI Development for Kids Act), aimed to regulate the use of children’s personal information by AI systems, Senate Bill 1018 (Automated Decisions Safety Act) regulating automated decision systems (ADS), and Senate Bill 420 (California AI Transparency Act).
August 4, 2025
Read our July AI Regulation Roundup to get the latest developments, announcements, and changes in the world of AI.
July 1, 2025
Read our June AI Regulation Roundup to get the latest developments, announcements, and changes in the world of AI.
June 1, 2025
Read our May AI Regulation Roundup to get the latest developments, announcements, and changes in the world of AI.
At Securiti, our mission is to enable organizations to safely harness the incredible power of Data & AI.
Copyright © 2025 Securiti · Sitemap · XML Sitemap
info@securiti.ai
Securiti, Inc.
3155 Olsen Drive
Suite 350
San Jose, CA 95117
