45 articles
OpenAI formally published a set of principles defining the conditions and limits of its collaboration with government entities and national security organizations, emphasizing democratic oversight and the protection of communities.
The European Commission today presented a joint Action Plan on Cybersecurity and Artificial Intelligence that builds on the existing EU legal framework, coordinates member states and industry, and addresses AI simultaneously as an attack vector and a defensive tool.
NVIDIA's AI Nations initiative shows how France, India, and Brazil are building domestic AI capacity according to a five-ingredient national strategy framework. The core thesis: AI data centers today are sovereign infrastructure as critical as roads and the power grid.
The UK AI Security Institute shows that AI agent evaluations with a fixed token budget systematically underestimate frontier capabilities. Scaling the budget from one to ten million tokens raises performance by up to 25 percent on software-engineering tasks and 22 percent on math benchmarks. AISI calls on regulators to move from benchmark scores to capability curves that account for variable compute budgets.
Mapping Europe's AI Workforce is an OpenAI report analyzing how AI technologies can reshape occupations and jobs across the European Union, with an emphasis on opportunities for the European workforce rather than exclusively on risks of automation.
The European Commission has issued a preliminary decision that AWS and Microsoft Azure should be designated as gatekeepers under the Digital Markets Act (DMA), which would subject both cloud giants to strict regulatory obligations to ensure fairer competition in the EU.
GitHub, Black Forest Labs, Hugging Face, and Mozilla have formed a coalition demanding amendments to the California AI Transparency Act (SB 942/SB 1000). A contested provision forces open-source projects to revoke licenses due to downstream user violations — incompatible with the perpetual license model.
OpenAI has joined Appia Foundation, a new international body developing shared evaluation standards and safety practices for advanced AI. The goal is global industry collaboration rather than fragmented national frameworks, and the announcement was made on June 23, 2026.
EUROPA is a consortium that the European Commission selected on June 19, 2026 as the winner of the Frontier AI Grande Challenge. Led by the Italian company Domyn, the consortium will build an open European frontier model with more than 400 billion parameters supporting all 24 official EU languages, to be made publicly available to companies, researchers and public institutions.
The European Commission published its fourth annual State of the Digital Decade report, assessing the EU's progress toward 2030 digital targets across infrastructure, business, skills, and public services. The report recognises AI, semiconductors, cloud, and open source as pillars of European technological sovereignty, and identifies delivery of results at scale, speed, and consistency as the key challenge. According to Eurobarometer, a large majority of Europeans rank digital policy at the top of EU priorities.
CNCF published a guide on architectural patterns for digital sovereignty of cloud-native platforms, as the EU Data Act has been fully applicable since January 11, 2025. Regulators expect four properties: jurisdictional boundedness, operational autonomy, cryptographic access control, and portability. A single Kubernetes cluster is insufficient because a shared control plane means shared risk; the proposed pattern gives each jurisdiction its own control plane (vCluster), with GitOps declaration of boundaries.
AgentFairBench is the first benchmark that measures demographic inequality in the actual actions of LLM agents — not just in their responses — across employment, lending, and medical triage domains. It uses counterfactual flip rate and action-rate disparity metrics and tests four agent scaffolds. In a pilot of 864 decisions, Claude Haiku showed no demographic effect above the noise floor, and the paper warns that naive comparison of six groups can overestimate inequality by roughly 2.4 times.
The paper introduces a Bayesian audit framework showing that a single final leaderboard snapshot of 1,000 systems can correspond to multiple incompatible historical trajectories, with convergence times ranging from 23 to 75 steps. Drawing on archived data from LiveBench, Open LLM Leaderboard v2, LMArena, GAIA, and tau-bench, the author proposes an archive-and-adjudication protocol for reconstructing scoring history and rejecting unsubstantiated claims about frontier models.
Anthropic has released a detailed regulatory framework proposal for frontier AI, introducing thresholds based on compute and company revenue, four categories of catastrophic risk, and requirements for independent evaluation and model protection.
The OECD analyzes how artificial intelligence can strengthen global food systems under pressure from climate, fragile supply chains, and labor shortages. The text highlights three barriers — siloed data, poor accessibility for smallholder farmers, and barriers to scaling. It recommends stronger international cooperation through the OECD and GPAI.
The OECD has published the first version of the AI Policy Toolkit, a practical and non-prescriptive guide that helps policymakers turn the OECD AI Principles into concrete action. The interactive platform with AI-powered semantic search follows the entire AI policy cycle, and it was created through multi-region co-creation with contributions from Italy, France, Korea, Japan, the UK and the European Union.
OpenAI has called for global action on youth safety in the context of artificial intelligence. The company proposes establishing an international institute that would strengthen safeguards, standards, and opportunities for young people, a policy proposal aimed at regulators and partners.
OpenAI published a shared playbook for external evaluations of AI models. The document describes how independent evaluators can reliably measure model capabilities, test safeguards and confirm the validity of results for advanced frontier systems.
The Frontier Governance Framework is a document published by OpenAI on May 28, 2026 that describes how its practices for AI safety, security and risk management align with the emerging regulations of the European Union and the state of California. The goal is to show the match between internal governance and external regulatory requirements.
The British AI Safety Institute (UK AISI) and the Australian AI Safety Institute signed an agreement on May 25, 2026, on cooperation in AI evaluation best practices and research findings exchange. The partnership expands the network of international AI safety institutions and strengthens a coordinated approach to risk assessment of advanced AI systems.
UK AI Safety Institute (AISI) published a report on 21 May 2026 analysing the future of oversight over advanced AI systems, based on 25 expert interviews from industry, government, and academia. The main finding: existing oversight rests on foundations that are likely to erode. More than 20 distinct degradation pathways for oversight mechanisms have been identified, with particular focus on latent reasoning, capability masking, external AI actions, and AI-to-AI communication.
The European Commission opened a targeted public consultation on 13 May 2026 on draft guidelines for classifying AI systems as high-risk under the EU AI Act. The consultation closes on 22 May at 18:00 CET, and the guidelines will directly determine which organisations in healthcare, education, critical infrastructure, and HR processes must meet the strictest regulatory requirements.
Google DeepMind has signed a national AI partnership with the Singapore government covering healthcare, education, and sustainability. By 2040, AI could contribute an additional $2.5 billion to the Singapore economy through accelerated R&D.
OpenAI is entering the second phase of its Education for Countries initiative — expanding partnerships with governments, launching the OpenAI Luminaries program for teachers, and offering certificates through OpenAI Academy. The goal is the systematic integration of AI tools into national education systems with measured real-world impact.
OECD.AI and the EU AI Office published an analytical report documenting how Europe is deploying artificial intelligence across four strategic sectors — agriculture, healthcare, industry, and mobility — with concrete active projects and identified barriers.
The European Commission has opened a public consultation on draft guidelines defining when an AI system falls under the "high-risk" category of the AI Act. Stakeholders — from developers to civil society — can submit feedback until June 23, 2026 at 10 PM CET.
How fast is autonomous AI cyber capability advancing? is a new report from the UK AI Safety Institute (AISI) published on May 13, 2026. Measuring cyber time horizons benchmarks (2.5M token budget, 80 % success threshold), AISI determined that the length of cyber tasks AI models can autonomously solve doubles every 4.7 months. Claude Mythos Preview is the first model to solve both cyber ranges (The Last Ones 60 %, Cooling Tower 30 %); GPT-5.5 solved The Last Ones at 30 %.
Anthropic Research 2028 AI Leadership Scenarios is a new policy paper published on May 14, 2026, describing two geopolitical scenarios for AI dominance by 2028. Scenario 1: US democracies maintain a 12-24 month lead through export controls and model defense. Scenario 2: China reaches parity through distillation attacks and $2.5B Supermicro-style chip smuggling. Anthropic recommends closing loopholes, defensive legislation, and championing American AI exports.
The Fine-Tuning FLOPs Meter toolkit is a new AWS SageMaker AI extension released on May 12, 2026, that automatically tracks the compute thresholds of the European AI Act (3.3×10²² FLOPs, 3.3×10²⁴ for systemic risk) during LLM fine-tuning. It is activated with a single flag compute_flops=true in the recipe YAML and automatically generates audit documentation to S3 and DynamoDB.
The European Commission has published draft guidelines on AI Act transparency obligations and opened a public consultation. The consultation runs until 3 June 2026, with obligations taking effect on 2 August 2026. Providers must label AI-generated content with machine-readable markers and notify users when they interact with AI systems.
The European Commission, Parliament and Council reached a political agreement on the Digital Omnibus package, which simplifies the AI Act and introduces an explicit ban on nudification apps. High-risk AI systems will apply from 02.12.2027, and AI integrated into products from 02.08.2028.
A new paper shows that safety in clinical LLMs does not follow the same scaling laws as accuracy — cleaner evidence in RAG raises accuracy from 73.5% to 94.1% and reduces high-risk errors from 12% to 2.6%, more than any model scaling effect.
The UK AI Security Institute announced a partnership with Microsoft on May 5 covering frontier AI safety. The collaboration spans three research areas: evaluation of high-risk capabilities, testing of safeguards, and research into societal resilience to conversational AI.
On May 5, 2026, NIST's Center for AI Standards and Innovation (CAISI) signed expanded agreements with Google DeepMind, Microsoft and xAI for pre-deployment and post-deployment testing of frontier models. CAISI has now conducted more than 40 evaluations, including unreleased state-of-the-art models, with testing routinely performed in classified environments with safeguards removed.
LangChain has published how LangSmith and LangChain OSS cover key articles of the EU AI Act — from risk management (Art. 9) to post-market monitoring (Art. 72). The deadline for high-risk AI systems is August 2, 2026, and penalties reach €15 million or 3% of global annual revenue.
On April 27, 2026, OpenAI announced FedRAMP Moderate authorization for ChatGPT Enterprise and OpenAI API. This opens the door to secure adoption of generative AI technology within US federal agencies, at a compliance level covering sensitive but unclassified data.
Chris Schneider, Philipp Schoenegger and Ben Bariach published on April 23, 2026 the paper 'Separable Expert Architecture' that solves one of the biggest GDPR problems of personalized LLMs: how to delete individual user data without retraining the entire model. The three-layer architecture (static base, composable LoRA adapters, per-user proxy artifacts) turns unlearning into a deterministic delete operation. Evaluated on Phi-3.5-mini and Llama-3.1-8B.
Google DeepMind and South Korea's Ministry of Science and ICT (MSIT) announced on 27 April 2026 a partnership to accelerate scientific discovery and AI innovation. The collaboration includes establishing an AI Campus in Seoul, providing academic institutions with access to advanced AI tools such as AlphaFold and AlphaGenome, and working with the Korean AI Safety Institute on safety issues.
The European Commission opened the DIGITAL-2026-BESTUSE-AWARENESS call worth 6 million euros to build a joint research framework against manipulative AI content. The call delivers the European Democratic Shield announced in November 2025 and accepts applications until October 1, 2026.
The European Commission has opened seven calls totaling €63.2 million through the Digital Europe Programme. The funding targets AI innovations in healthcare (cancer, heart disease), online child safety and tools for regulators, and forms part of the broader AI Continent Action Plan.
AISafetyBenchExplorer is a structured catalog documenting 195 AI safety benchmarks published between 2018 and 2026. The research reveals alarming fragmentation in the field — terms like 'accuracy' and 'safety score' conceal entirely different methodologies. Of the 195 benchmarks, 165 evaluate only the English language, and 137 have inactive GitHub repositories, indicating a lack of maintenance after publication.
The OECD analyzes the UK's Algorithmic Transparency Recording Standard (ATRS), mandatory for central government since 2025. By March 2025, 125 records on algorithm use have been published. Estonia has already adopted the standard, and the OECD calls it 'world-leading.'
Anthropic has announced the appointment of Vas Narasimhan, CEO of Novartis, to the board of directors through the Long-Term Benefit Trust. LTBT directors now constitute a board majority, strengthening oversight of the company's safety mission.
Researcher Tibebu proves a formal impossibility result: above a certain threshold of AI agent autonomy, all four properties of accountability cannot simultaneously hold in systems combining humans and AI.
The European Commission has marked one year since launching the AI Continent Action Plan by publishing the results: 19 AI factories have been deployed across European supercomputers, 13 regional AI Factory antennas provide local access, and one billion euros has been secured for Apply AI Strategy initiatives.