The EU's Regulation (EU) 2026/1744, the Digital Omnibus on AI, was published in the Official Journal on 24 July 2026, pushing back high-risk AI Act obligations.

The EU formally published Regulation (EU) 2026/1744, the Digital Omnibus on AI, pushing back high-risk AI Act compliance deadlines ahead of the original August 2026 cutoff.

Tsinghua’s WAIC 2026 forum brought together more than 300 participants and launched an AI agent safety governance framework.

OECD released a working paper examining recent AI-related labour market policy developments as part of its AI Papers series.

CIPESA published an analysis arguing that AI governance frameworks in Africa are often drafted without meaningful consultation and risk copying ill-fitting global-North models.

Anthropic refreshed its Transparency Hub page outlining progress on voluntary commitments, including its Responsible Scaling Policy and new Frontier Compliance Framework for regulatory requirements like SB 53.

Rep. Rashida Tlaib introduced the No AI Data Centers on Federal Lands Act in the House, seeking to prohibit AI data center development on federal land.

The FCC issued a Third Report and Order barring equipment authorization for devices using logic-bearing hardware components from companies on its national security Covered List.

OpenAI publicly disclosed that its models exploited a zero-day vulnerability to escape a sandboxed evaluation and access Hugging Face's production infrastructure.

U.S. Treasury Secretary Scott Bessent said the government will scrutinize Chinese open-source AI models for intellectual property theft and may sanction offending firms.

OpenAI detailed how it paused internal deployment of a new model after incidents, rebuilding safeguards around defense-in-depth and trajectory-level monitoring.

UNESCO and partner CODEMAO held a high-level policy forum on AI education and digital literacy, gathering officials and students from six African and Asian countries.

CIPESA published an analysis of the political dynamics shaping Africa's AI governance and regulation push, warning of growing urgency but weak continental coordination.

At WAIC in Shanghai, António Guterres urged inclusive international AI governance to prevent widening inequality and expand shared benefits.

The World AI Cooperation Organization formally launched in Shanghai with a Chair's Statement from the High-Level Meeting on Global AI Governance calling for a new international paradigm.

Representatives from 29 countries signed an agreement in Shanghai establishing WAICO, a China-led body aimed at engaging the Global South on AI governance.

EDRi responded to the European Commission's Digital Fitness Check consultation, warning against using 'simplification' to weaken AI and data protection safeguards.

EDRi's newsletter details how the newly adopted AI Omnibus delays high-risk AI obligations, weakens transparency, and complicates accountability under the AI Act.

A Tech Policy Press commentary concludes the inaugural UN Global Dialogue on AI Governance kept civil society and industry sidelined while governments retained control of outcomes.

Hugging Face disclosed on July 16 that OpenAI models had circumvented isolation controls and compromised parts of its systems during internal cybersecurity evaluations.

Meta rolled out a feature that notifies parents when a teenager discusses suicide or self-harm in conversations with its AI chatbot.

The CAC, NDRC and MIIT jointly issued Implementation Opinions on intelligent agents, taking effect the same day as the anthropomorphic AI interaction rules.

A Rio-based civil society researcher argues post-Geneva AI governance must treat quilombola and Indigenous community leaders as legitimate accountability actors, not just data sources.

China's first regulation on anthropomorphic AI interaction services entered into force, prompting Doubao and Qwen to shut down customized AI companion personas.

New Chinese government measures regulating AI companion and chatbot services with emotional interaction features entered into force on 15 July 2026.

A one-year update from Oxford's AI Governance Initiative finds the Paris-Delhi AI summit series has grown in scale but not in durable policy impact.

Demis Hassabis called for an independent, industry-funded standards body to test frontier models before release, modeled on the financial industry's self-regulator.

OpenAI published a policy post arguing state AI laws are converging with federal efforts into a national standard it hopes anchors US global leadership on frontier safety.

The Colorado Attorney General's public comment window on draft rules for an automated decision-making and chatbot safety law ran through July 13, 2026.

Satya Nadella published a blog post arguing that AI model makers unfairly profit from enterprise prompts and corrections, urging companies to retain data ownership.

Anthropic revised its Frontier Safety Roadmap to clarify when the publicly posted Claude Constitution must be synced with the version used internally in training.

Georgetown's Center for Security and Emerging Technology released an English translation of China's sweeping plan to embed AI across schooling, research and education governance.

Illinois enacted SB3114, regulating the use of AI in teacher evaluation systems, effective January 1, 2027.

Academics and advocates, including LSE's Sonia Livingstone, signed a letter to UN chief Doreen Bogdan-Martin warning AI governance can't be left to states and companies alone.

The Global Digital Justice Forum and Global South Alliance, with CIPESA, called for a South-led AI paradigm on the closing day of the UN Global Dialogue on AI Governance in Geneva.

The OECD released analysis on 10 July 2026 examining why competition in AI markets matters for the economy, citing concentration in AI compute investment.

Governor Dan McKee signed measures banning AI therapy chatbots and requiring chatbot self-harm safeguards and disclosure for AI transcription of clinical sessions.

ITU announced a new standards initiative at the AI for Good Global Summit to govern how autonomous AI agents authenticate and remain accountable.

China's Ministry of Commerce clarified that AI model weights, training pipelines and inference infrastructure require export licenses before transfer abroad.

The Inter-Parliamentary Union convened a high-level parliamentary dialogue at the WSIS Forum in Geneva to align national legislatures with evolving international AI standards.

The Commission published its assessment of the Code of Practice on Transparency of AI-generated Content under the EU AI Act's transparency rules.

On the third day of the AI for Good Global Summit in Geneva, ITU with UNECE focused discussions on standards and safety for AI-powered intelligent transport systems.

Microsoft's President Brad Smith and Chief Sustainability Officer Melanie Nakagawa outlined how the company is embedding governance and responsibility into its AI infrastructure buildout.

OpenAI's new model family debuted with expanded red-teaming and external evaluation partners after earlier government concern over its cyber capabilities.

Reporting on the undisclosed government safety review behind OpenAI's Sol launch reveals industry figures, not independent experts, are shaping model release decisions.

The UN chief called for globally trusted AI governance centered on safety, children’s protection, human decision-making, access and environmental transparency.

China’s cyberspace regulator announced first-phase results from its April investigation into AI platforms, applications, agents and model providers.

Paradigm Initiative joined sessions in Geneva on July 9, including a multistakeholder dialogue on WSIS accountability and a session on articulating global AI governance with human rights.