Other
AI: capabilities, regulation, labour
Updated 6h ago·Important
Models are learning faster than the institutions behind them can adapt. The thread tracks frontier-model capability jumps, the AI Act and its enforcement, labour-market impact, and infrastructure (chips, energy, water).
Latest update
AI Office launches first systemic-risk probe into frontier model
The European Commission's AI Office opens its first formal investigation into a frontier general-purpose AI model for systemic risk, citing new benchmark evidence of rapidly advancing autonomous cyber operations capabilities.
State of play
The EU's regulatory machinery is now in motion, moving from a single landmark investigation to a multi-front enforcement campaign. The AI Office's first systemic-risk probe into a frontier model's autonomous cyber capabilities is no longer an isolated test case. It is the spearhead of a coordinated strategy, backed by newly operational national safety institutes in Germany, France, and Spain, and informed by the Commission's first technical guidance on testing for such risks. This top-down pressure on frontier models is matched by a bottom-up crackdown, as labour inspectors across four member states simultaneously target opaque workplace AI tools, issuing the Act's first tangible fines. Parallel legal battles over training data transparency are expanding, with publishers now demanding court-appointed audits. Meanwhile, the physical limits of the AI buildout are becoming stark, as chipmakers warn that European energy grids cannot support planned fabs, and Nordic authorities scrutinise the surging water use of AI data centres. The bloc's layered response is taking shape, but its ultimate speed is challenged by accelerating model capabilities and concrete infrastructure bottlenecks.
Chronicle
View historyCoordinated EU enforcement targets workplace AI systems
National authorities in multiple EU member states begin coordinated enforcement actions against AI-powered hiring and worker-monitoring tools, applying the AI Act's high-risk rules and demanding compliance evidence.
German and French AI safety institutes begin active testing
Germany and France's new AI safety institutes transition from setup to active technical work, conducting structured evaluations and red-team exercises on frontier models to support EU and national regulatory assessments.
Benchmarks show rapid gains in AI cyber autonomy
Recent benchmark reports indicate frontier models' autonomous cyber and software engineering capabilities have roughly doubled in complexity over a timescale of months, directly informing government risk assessments.
EU AI Office opens first systemic-risk investigation
The European Commission's AI Office launches its first investigation under the AI Act's systemic-risk provisions, targeting a major frontier foundation model. The probe focuses on the system's ability to autonomously conduct offensive cyber operations with minimal human oversight, examining whether required testing and documentation were completed.
National safety institutes begin coordinated frontier testing
Germany's BKAI, France's national institute, and Spain's Observatorio de Seguridad de la IA start running structured evaluations on frontier models in coordination with the EU AI Office. The shared testbeds focus on autonomous cyber and CBRN-related capabilities, aiming to prevent duplication and build a common evidence base for EU enforcement.
EU labour authorities launch coordinated crackdown on workplace AI
Inspectors in Italy, the Netherlands, Sweden, and Portugal open simultaneous investigations into AI-based productivity and surveillance tools in sectors like logistics and call centres. Initial fines and corrective orders have been issued in Italy and the Netherlands for deploying tools without proper risk assessments or worker consultation.
Publishers expand lawsuits to demand training data audits
Publishers and collecting societies in Germany, Italy, and the Netherlands amend existing copyright suits against OpenAI and Anthropic. The new complaints demand court-appointed technical experts to audit the models' training data to verify compliance with EU copyright law and the AI Act's transparency rules.
Nordic regulators scrutinise AI data centre water use
Authorities in Denmark and Finland open reviews into the environmental impact of hyperscale data centres used for AI training, focusing on sharply increased water consumption for cooling during heatwaves. The probes may lead to tighter reporting rules or dynamic pricing for peak water use.
Commission circulates first guidance on testing cyber capabilities
The European Commission issues non-binding technical guidance to national authorities and model providers, outlining baseline tests for identifying systemic-risk autonomous cyber capabilities. The document recommends red-teaming for end-to-end exploit chains and measuring how much a model accelerates novice attackers.
Chipmakers warn European energy grids threaten AI fab plans
Executives from TSMC and Intel tell European policymakers that grid constraints and permitting delays risk slowing planned chip manufacturing expansions in Germany, Ireland, and Poland. The warnings highlight a growing tension between AI infrastructure ambitions and the physical capacity of European energy networks.
Anthropic wzywa do globalnego wstrzymania rozwoju AI, przygotowując się do IPO i naprawiając relacje z Białym Domem
Firma AI stojąca za Claude'em twierdzi, że zaawansowane modele mogą wkrótce samodzielnie się ulepszać bez ludzkiej kontroli, i proponuje skoordynowane globalne wstrzymanie. Apel pojawia się, gdy Anthropic poufnie składa dokumenty IPO i pracuje nad naprawą relacji z administracją Trumpa.
Rozpoczyna się roadshow IPO SpaceX; S&P 500 blokuje wcześniejsze wejście do indeksu, a dostęp z Chin jest ograniczony
SpaceX Elona Muska rozpoczął marketing swojej oferty publicznej o wartości 75 miliardów dolarów w Nowym Jorku, ale S&P 500 odmówił przyspieszenia jej wejścia do indeksu, a inwestorzy z Chin i Hongkongu zostali zablokowani w dostępie do dokumentów oferty.
Anthropic wzywa do globalnego zamrożenia AI, naprawiając relacje z Białym Domem przed IPO
Anthropic, gigant AI z San Francisco wyceniany na około bilion dolarów, wezwał w czwartek do globalnie skoordynowanego wstrzymania prac nad granicznymi modelami sztucznej inteligencji, ostrzegając, że samodoskonalące się modele mogą wymknąć się spod ludzkiej kontroli. Propozycja pojawia się, gdy firma łagodzi wielomiesięczny spór z administracją Trumpa i przygotowuje się do głośnego debiutu giełdowego.
Anthropic wzywa do globalnego wstrzymania zaawansowanego rozwoju AI, powołując się na ryzyko rekurencyjnego samodoskonalenia
Amerykańska firma AI Anthropic zaproponowała globalny mechanizm koordynacji w celu spowolnienia lub zawieszenia rozwoju systemów AI na granicy możliwości, argumentując, że społeczeństwo potrzebuje czasu na dostosowanie się do szybkiego postępu technologicznego.
Industry consortium advances capability testing for high-risk AI behaviors
Technical reports from the Frontier Model Forum detail the industry's growing focus on pre-deployment capability assessments. These evaluations aim to measure specific high-risk behaviors, such as autonomous cyber and biological capabilities, moving beyond reliance on model size or training compute. This industry-led approach aligns with the regulatory trend toward benchmark-based oversight.
Policy analysis calls for dynamic, capability-based frontier AI definitions
A policy memo from Third Way argues that the definition of frontier AI must remain dynamic and capability-focused. It states that static compute thresholds are merely a proxy and that regulators require the discretion to assess risks on a case-by-case basis as models evolve. The analysis notes this discretionary power is already embedded within the EU AI Act's framework for systemic risk.
First EU systemic-risk probe tests AI Act's dynamic enforcement
The European Commission opens its first formal investigation under the AI Act's systemic-risk provisions, targeting a major frontier model provider. The probe will examine safety testing, incident reporting, and transparency, and may set a precedent for using capability metrics to classify models as 'frontier' even below the Act's compute threshold.
Germany and France launch national AI safety bodies
Germany and France establish national AI safety institutes, aligning with the EU AI Act. The German institute will red-team advanced models, while France expands its digital regulator's remit. Both bodies are set to collaborate with the EU AI Office and the UK's AI Safety Institute on shared evaluation protocols.
US and EU coordinate on tighter AI chip export controls
The US tightens export controls on high-end AI accelerators and lithography tools, with the EU assessing complementary measures. Officials link these hardware controls to compute-based regulatory thresholds, viewing them as a necessary backstop while capability-based rules are operationalised.
AI-linked layoffs spread across European sectors
Major layoffs linked to generative AI adoption are reported across European media, IT services, and back-office sectors. Company statements describe a pattern of augmentation, hiring freezes, and restructuring, with losses concentrated in junior and routine positions. This intensifies calls for stronger just-transition measures.
Autonomous AI cyber prowess now doubles every five months
New analysis from the UK AI Safety Institute finds the maximum length of autonomous cyber tasks frontier models can complete at 80% reliability is doubling roughly every 4.7 months. This acceleration is reshaping software and cybersecurity labour markets, compressing tasks that took a year into weeks.
Industry expands frontier model tests to match regulatory regimes
An industry consortium of AI developers releases an updated frontier capability assessment framework. It integrates autonomous-cyber, tool-use, and long-horizon planning tests designed to mirror requirements from the EU AI Act and UK AI Safety Institute, including red-teaming for critical-infrastructure risks.
AI's rising energy and water demand flagged as systemic risk
European infrastructure planners warn that the rapid expansion of AI data centres is straining regional power grids and water supplies. Several member states are considering requirements for energy-efficiency benchmarks, waste-heat reuse, and location constraints for new projects.
Copyright suits against AI firms advance in European courts
Copyright and training-data lawsuits against OpenAI and Anthropic gain traction in EU courts, with plaintiffs arguing their practices breach EU rules. The cases intersect with new AI Act obligations for training-data summaries, potentially shaping how firms document datasets and negotiate licensing in the EU.
EU opens first systemic-risk investigation under AI Act
The European Commission has initiated its first formal investigation into potential systemic risks posed by two unnamed frontier general-purpose AI models. The EU AI Office and DG CONNECT are leading the probe, requesting detailed documentation on training compute, safety evaluations, and deployed safeguards. Officials frame the inquiry as a template for future enforcement, indicating that capability-based criteria may be used to designate systemic-risk models even below the formal compute threshold. This marks the concrete shift of the AI Act from abstract obligations to active supervision.
Germany and France launch national AI safety bodies
Germany has formally inaugurated a Federal AI Safety Institute in Berlin to provide technical evaluations and red-team testing for high-risk systems. France has announced a national AI risk observatory tasked with monitoring safety and societal risks from large models. Both bodies are mandated to work closely with the EU AI Office and national sector regulators, supplying independent assessments and feeding evidence into enforcement actions. Their creation aligns with similar institutes in other member states, building a distributed technical capacity to support the AI Act's enforcement across the bloc.
Autonomous cyber capability advances on five-month doubling cycle
Analysis by the UK's AI Safety Institute finds the length of cyber tasks frontier models can autonomously complete at high reliability has been doubling roughly every four to five months since late 2024. European cybersecurity agencies warn these advances could outpace existing defensive postures, especially for smaller organisations. The capability to chain software vulnerabilities and conduct targeted campaigns at scale is now observed in red-team exercises. This rapid progress is pushing EU officials to consider incorporating cyber-offence benchmarks into systemic-risk assessments.
Industry expands evaluation suites ahead of EU scrutiny
Major AI developers are publishing expanded evaluation suites for new frontier models, adding tests for autonomous software exploitation, biosecurity reasoning, and long-horizon planning. These disclosures are explicitly framed as preparation for AI Act obligations on high-impact models, which will require continuous monitoring and risk reporting. Industry groups are collaborating with external labs to define capability thresholds that could trigger stricter safeguards. European regulators welcome the transparency but stress that internal evaluations will not substitute for independent testing.
Data centre regulations tighten over AI energy and water use
Member states including the Netherlands, Spain, and Ireland are introducing or proposing tighter permitting and reporting requirements for large data centres hosting AI workloads. The moves are driven by concerns over surging electricity demand and water consumption for cooling. Some AI-linked projects have been delayed or conditioned on implementing efficiency measures or heat-reuse schemes. Grid operators warn that rapid AI compute growth risks localised capacity constraints, creating a new intersection between digital regulation and environmental compliance.
EU launches first AI Act systemic-risk investigation
The European Commission opens its first formal systemic-risk investigation under the AI Act, targeting two unnamed frontier model providers. Coordinated with Germany's AIDA and France's INRIA-based safety institute, the probe will involve red-team exercises focused on risks of autonomous cyber and critical-infrastructure misuse.
Commission prepares first fines for AI Act violations
Brussels sends formal notices of violation to several general-purpose AI providers for failing to meet early transparency and cybersecurity obligations, including missing training-data summaries and incident reports. Officials are preparing the first fine proposals.
National AI safety institutes formalise cross-border coordination
Germany's AIDA and France's AI security institute, now formally mandated for AI Act enforcement, deepen coordination with the UK AI Safety Institute. They agree on shared evaluation benchmarks and joint-testing protocols for frontier models, pooling technical resources.
Tech sector restructuring hits European labour market
Several large technology and media companies announce AI-linked layoffs in Europe, cutting back-office and support roles while expanding hiring for AI engineering and governance positions. Trade unions press for stronger consultation on AI-driven reorganisation.
Benchmarks confirm rapid gains in autonomous cyber capabilities
Updated evaluations by safety institutes show frontier models can reliably complete longer autonomous cyber tasks, with the 80% reliability horizon doubling every 4 to 5 months. This pace compresses the response time available to regulators.
EU launches first systemic-risk probes under AI Act
The European Commission has opened its first formal investigations into two unnamed frontier AI model providers for potential systemic risks. The probes, coordinated with new national institutes in Germany and France, could lead to strict obligations on testing, red-teaming, and incident reporting.
Commission sends first AI Act violation notices
In a parallel enforcement action, the Commission issued its first formal notices of violation to several model providers for missing deadlines on publishing training-data summaries, copyright policies, and security documentation.
National AI safety institutes become operational
Germany and France have formally launched their national AI safety institutes. These bodies are tasked with technical evaluations and red-teaming, and are already supporting the Commission's systemic-risk investigations.
Benchmarks confirm rapid gains in autonomous cyber capabilities
New evaluations show state-of-the-art frontier models have sharply improved at autonomous cyber operations since early 2026, with the effective length of tasks they can complete continuing to double every few months.
EU debates tightening AI chip export controls
EU trade and foreign ministers have opened discussions on whether to tighten controls on exports of advanced AI chips and manufacturing equipment, following recent moves by the United States and Japan.
AI-driven corporate restructuring accelerates in Europe
Several large European firms in finance, telecoms, and retail have announced restructuring plans explicitly citing generative AI deployment as a factor in reducing or reshaping white-collar roles.
Copyright and GDPR probes into AI training broaden
Supervisory authorities and courts in multiple EU member states have opened or expanded investigations into how large model providers use copyrighted works and personal data for training.
UK and EU AI safety institutes increase coordination
The UK's AI Safety Institute and emerging EU national institutes have stepped up technical collaboration on evaluating frontier models, sharing workstreams on benchmarking autonomous cyber capabilities and agentic behaviour.
Germany and France establish national AI safety institutes for enforcement support
Germany has formally established a federal AI safety institute in Bonn, the Kompetenzzentrum für KI‑Sicherheit, with a mandate to test high-risk and frontier models to support AI Act enforcement. France has similarly tasked its new Paris-based Institut français de l’IA with regulatory-support functions, including red-team design and evaluation protocols. Both bodies will provide the technical foundation for the European Commission's ongoing systemic-risk and violation investigations.
EU institutes and UK AI Safety Institute align on autonomous cyber benchmarks
The newly formed German and French institutes have signed a memorandum of understanding with the UK's AI Safety Institute. The pact aims to align a core set of benchmarks, avoiding duplicated testing, with initial joint work focusing on standardising evaluations for models capable of multi-step software exploitation and automated vulnerability discovery.
UK institute reports frontier model cyber autonomy accelerating faster than projected
The UK AI Safety Institute's updated analysis shows the length and complexity of cyber tasks frontier models can complete autonomously has been doubling roughly every 4–5 months since late 2024. This pace has shortened the 80%-reliability 'time horizon' for autonomous cyber operations into 2026, outpacing previous late-2025 projections. The institute warns this increases feasibility of models chaining long-horizon intrusion and lateral-movement tasks.
Anthropic calls for coordinated global development pause citing self-improvement risk
Anthropic, the firm behind Claude, has publicly called for a coordinated global pause on advanced AI development, arguing models may soon be capable of self-improvement without human control. The appeal comes as the company privately files IPO documents and works to mend relations with the new US administration.
EU AI Office opens first systemic-risk investigations
The European Commission's AI Office has initiated its first formal investigations into several unnamed frontier AI models, designating them as potential systemic risks under the AI Act. The probes focus on autonomy in cyber operations and large-scale manipulation harms, leveraging new powers to demand technical documentation and order evaluations. Officials cited rapidly advancing agentic capabilities, as highlighted by UK and independent benchmarks, as the key justification.
Germany and France formally establish AI safety institutes
Germany's Bundesinstitut für KI-Sicherheit (BIFAS) and France's Institut français de l'IA de confiance are now formally operational. Both have signed technical cooperation agreements with the UK AI Safety Institute to exchange evaluation methods and share red-team results on high-risk capabilities like autonomous cyber operations. The institutes will support AI Act enforcement but not replace enforcement decisions by supervisory authorities.
National authorities coordinate first labour-market AI Act probes
Authorities in Italy, the Netherlands, and Spain have opened coordinated enforcement actions against AI-based workplace monitoring and algorithmic hiring tools. The cases target opaque productivity-scoring and CV-screening systems in logistics, retail, and public administration, focusing on discrimination, transparency, and human oversight. This marks the AI Act's first concrete application to labour-market uses, integrating national labour inspectorates into the supervisory structure.
New collective lawsuits target training data transparency
Consumer and creator organisations in Italy, Germany, and the Netherlands have filed new collective actions against OpenAI and Anthropic. The lawsuits allege unlawful use of copyrighted works and personal data in training, explicitly invoking the AI Act's transparency and data-governance provisions for systemic-risk providers. Claimants seek detailed training-data summaries, opt-out mechanisms, and remuneration schemes, testing the interaction between copyright, GDPR, and the new AI law.
Joint EU task force formed on AI-cyber threats
The EU AI Office, ENISA, and Europol have established a joint task force to monitor and respond to cyber threats amplified by frontier AI models, including automated vulnerability discovery. The group will coordinate with national cybersecurity teams and the UK institute to share red-team results and draft guidance on security obligations for systemic-risk model providers under the AI Act, reflecting concerns that attack capabilities are outpacing institutional responses.
First AI Act Systemic-Risk Probe Launched
The European Commission formally launches its first systemic-risk investigation under the AI Act, targeting a frontier model with advanced autonomous cyber capabilities. The probe will test the regime's ability to compel developers to provide post-deployment evaluations and red-teaming data.
Italy and Spain create joint taskforce on AI labour impact
Italy and Spain form a joint taskforce on AI and employment to coordinate responses to automation-driven restructuring. The taskforce will collect data on job impacts and explore EU-level funding for reskilling, as unions push for consultation clauses in collective bargaining.
First systemic-risk probe targets two frontier models
The European Commission launches its first formal investigation under the AI Act's systemic-risk provisions, targeting two unnamed frontier models. The probe, coordinated by the EU AI Office, will assess compliance with risk management, cybersecurity, and incident reporting obligations.
Capability-based triggers fast-tracked into AI Act rules
In parallel, the Commission begins revising the AI Act's delegated acts to embed capability-based triggers, such as autonomous cyber horizons, as the primary criteria for defining systemic-risk models. This move aims to make regulatory thresholds explicitly adjustable based on model evaluation data.
Germany launches federal AI safety lab
Germany establishes a federal AI Safety and Security Lab under its Federal Office for Information Security (BSI) to test high-risk and systemic-risk models. The lab will collaborate with the EU AI Office, focusing on models used in industrial control, energy, and healthcare.
France opens national AI evaluation centre
France inaugurates a National AI Evaluation Centre tasked with standardized safety testing for both large foundation models and smaller systems embedded in public-sector use cases. It will support regulators and participate in joint testing with international partners.
Nordic agreements pivot AI rollout to redeployment
New sectoral agreements in Sweden, Denmark, and Finland commit large employers to prioritize retraining and role changes over layoffs when introducing advanced AI. The deals include joint monitoring committees to review productivity and employment effects.
UK Safety Institute confirms acceleration in autonomous cyber capabilities
The UK AI Safety Institute publishes updated longitudinal data, confirming that the doubling time for the complexity of autonomous cyber tasks frontier models can reliably complete remains at or below 4.7 months. The institute's analysis of models released in early 2026 shows the trend holding as context windows and tool-use improve.
First AI Act Systemic-Risk Probes Begin
The European Commission launches the first formal systemic-risk investigations under the AI Act into two unnamed frontier model developers. The probes test the Act's new powers to demand post-deployment logs, incident reports, and red-team data to verify compliance with continuous monitoring and emergent-risk reporting obligations.
Germany, France Launch National Safety Institutes
Germany and France establish national AI safety institutes with mandates to adversarially test frontier models. Germany's Bundesinstitut für KI‑Sicherheit will focus on models in critical infrastructure, while France expands its Paris AI lab into a national evaluation hub, both aiming to feed results into the EU's systemic-risk classification process.
Autonomous Cyber Capability Growth Confirmed
New data from the UK AI Safety Institute confirms the complexity of autonomous cyber tasks frontier models can complete continues to double roughly every five months. Independent benchmarks show these models can now compress year-long manual penetration tests into weeks, executing multi-step intrusion chains autonomously.
EU Proposes Dynamic Risk Test to Replace FLOP Threshold
The Commission's AI Office circulates a draft implementing act proposing to replace the AI Act's static compute threshold with a dynamic risk test based on real-world capability benchmarks, such as autonomous cyber offence. The shift is explicitly driven by the empirical acceleration timelines published by safety institutes.
Brytyjski regulator nakazuje Google umożliwienie wydawcom rezygnacji z AI w wynikach wyszukiwania — światowy precedens
Brytyjski Urząd ds. Konkurencji i Rynków nałożył na Google nowe wymogi dotyczące postępowania, zmuszając giganta technologicznego do umożliwienia wydawcom blokowania ich treści przed funkcjami wyszukiwania opartymi na sztucznej inteligencji, bez uszczerbku dla ich pozycji w standardowych wynikach wyszukiwania.
UK safety institute reports renewed acceleration in autonomous cyber capabilities
The UK AI Safety Institute releases new data indicating the effective time horizon for models to complete autonomous cyber tasks has shortened, with capabilities on complex software tasks doubling in roughly four months. The findings show models can sustain multi-step intrusion sequences with limited oversight.
EU Commission signals readiness to adjust AI Act thresholds
Senior European Commission officials state they are prepared to revise the AI Act's frontier-model thresholds and systemic-risk designations. The AI Office is studying whether current compute thresholds and testing templates are sufficient to capture powerful autonomous-agent behaviours, signalling a move toward more capability-based criteria.
Commission moves to tighten frontier-model thresholds
The European Commission begins drafting a delegated act to formally update the AI Act's definition of 'systemic risk' frontier models. The planned change would shift the framework toward more dynamic, capability-focused criteria, allowing regulators to impose obligations on models with advanced autonomous capabilities, even if their training compute falls below the original 10^25 FLOP threshold.
France and Germany establish national safety institutes
Germany advances legislation to create a federal AI Safety and Security Institute, while France expands its national AI hub into a dedicated frontier model evaluation centre. Both institutes are mandated to conduct independent safety testing and red-teaming, feeding technical evidence into the EU AI Office's systemic-risk assessments.
Transatlantic talks begin on AI chip export controls
US and EU policymakers begin discussions on coordinating tighter export controls for advanced AI chips, aiming to restrict access to high-end hardware for training frontier models in countries deemed high-risk for military or cyber misuse.
EU opens first systemic-risk investigation under AI Act
The European Commission initiates its first formal probe into a frontier AI model designated as posing a potential systemic risk. The investigation focuses on whether the model's autonomous cyber and code-generation capabilities breach obligations for risk monitoring and mitigation, testing the AI Office's powers to demand post-deployment evaluations and incident reports.
Technical data quantifies rapid acceleration in autonomous cyber capabilities
Analysis from the UK AI Safety Institute estimates the length of autonomous cyber tasks frontier models can complete has been doubling roughly every 4 to 5 months since late 2024. This data is cited by EU officials as concrete evidence that fixed compute thresholds are obsolete, bolstering the case for dynamic, capability-based regulatory triggers.
Security research underscores near-term autonomous threat
Reports from security researchers and the Frontier Model Forum indicate current models can autonomously chain exploits and adapt to defenses, completing vulnerability discovery work comparable to a year of manual penetration testing in under three weeks. This reinforces policy arguments for capability-based risk designations.
Regulators advance capability-based definitions for frontier AI
Policy memos describe a clear shift in the EU and some US states toward defining systemic-risk AI by demonstrated capabilities like advanced cyber functions, moving beyond static training-compute thresholds. The EU is preparing secondary legislation to operationalize this flexible approach within the AI Act.
UK competition regulator sets precedent on AI and publisher content
The UK's Competition and Markets Authority orders Google to allow publishers to opt their content out of AI-powered search features without penalty to their ranking in standard search results. This creates a global precedent for separating AI training and indexing from core platform services.
EU AI Office launches first systemic-risk investigation
The European Commission's AI Office has initiated its first formal probe under the AI Act, examining a frontier-model developer's pre and post-release safety testing, capability evaluations, and incident reporting. National AI safety institutes are providing technical support, creating a test-bed for future EU enforcement workflows.
France inaugurates national AI safety institute
The French government has officially launched a national AI safety institute with a multi-year budget. Its mandate includes technical evaluation of frontier models, focusing on capabilities like autonomous cyber operations, to feed findings into the EU AI Office's systemic-risk assessments.
Germany scales federal AI safety centre for EU audits
Germany has expanded its federal AI safety and trust centre, giving it a formal mandate to conduct technical audits of systemic-risk models designated under the AI Act. The centre is recruiting specialists in robustness and cyber security to support EU-level capability evaluations.
Commission drafts EU-wide controls on AI chip exports
The European Commission is preparing a proposal for coordinated controls on exports of advanced AI chips and large training clusters, aiming to align with US and Japanese restrictions. The draft framework also explores licensing for large cloud-based compute allocations to non-EU entities.
AI Office issues guidance linking capabilities to risk designation
The EU AI Office has published guidance clarifying that systemic-risk designation can be based on demonstrated capabilities (like autonomous cyber operations) beyond just training compute. It outlines expectations for developer-supplied evaluation results, tying testing practices directly to compliance.
UK institute reports sustained rapid gains in AI cyber autonomy
The UK AI Safety Institute reports that the length and complexity of cyber tasks frontier models can complete autonomously continues to grow, with a doubling time for reliable tasks near 4–5 months. European regulators cite this as evidence for dynamic, capability-based oversight.
Microsoft Build 2026: chip kwantowy z celem na 2029, zapowiedź noszonych urządzeń AI i pogłębienie partnerstwa z Nvidią
Podczas konferencji deweloperskiej Build 2026 Microsoft zaprezentował nowy chip kwantowy z celem na 2029, zapowiedział koncepcje noszonych urządzeń zasilanych AI oraz pogłębił partnerstwo sprzętowe z Nvidią.
First systemic-risk probe launched under AI Act
The EU AI Office opens its first formal investigation into a frontier foundation model, using capability-based criteria like autonomous cyber performance to assess systemic risk. National authorities in three member states are called to assist with technical testing.
National AI safety institutes expand for EU Act
France and Germany announce expansions of their national AI safety institutes, explicitly aligning their mandates with the EU AI Act's needs for red-teaming, safety testing, and post-market monitoring of general-purpose models.
AI data centre boom strains EU infrastructure
Energy and water regulators across the EU warn that the projected boom in AI data centres is straining local power grids and water systems, especially in southern Europe. The Commission examines whether to update sustainability reporting rules for large data centres.
EU moves to tighten AI chip export controls
The European Commission proposes updated dual-use export controls, coordinated with the US and Japan, to require licenses for advanced AI accelerators and high-bandwidth memory destined for jurisdictions of concern. Negotiations begin over potential exemptions for domestic semiconductor firms.
Autonomous cyber capability growth accelerates
New analysis from the UK AI Safety Institute finds the time horizon for frontier models to perform complex, autonomous cyber operations is shortening faster than expected, with capabilities now doubling every 4.7 months. This acceleration outpaces previous projections.
AI automation cited in new wave of job cuts
Major employers in Germany, the Netherlands, and Spain announce white-collar job cuts linked to productivity gains from generative AI tools automating reporting, legal drafting, and customer support. Trade unions call for faster EU guidance on labour law in AI-driven reorganisations.
Copyright battles over AI training data intensify
Coordinated copyright lawsuits are filed or expanded in France, Germany, and Italy against OpenAI and Anthropic, alleging unlicensed use of protected text and images for training. The suits test the interaction of EU copyright law with the AI Act's new transparency duties.
Trump podpisuje okrojoną dyrektywę w sprawie AI, dając rządowi 30-dniowy dobrowolny przegląd zaawansowanych modeli
Prezydent Trump podpisał we wtorek zarządzenie wykonawcze, w którym prosi firmy z branży AI o udostępnianie swoich najpotężniejszych modeli do dobrowolnego przeglądu rządowego na maksymalnie 30 dni przed publiczną premierą, wycofując się z wcześniejszych projektów, które nakazywałyby 90-dniowe okno.
AISI Quantifies Explosive Acceleration in Autonomous AI Cyber Capabilities
The UK AI Safety Institute (AISI) publishes analysis showing the 'task-length horizon' for frontier models to autonomously complete cyber operations is doubling every 4.7 months, a pace comparable to the most explosive software task improvements. This acceleration, tracked since late 2024, places cyber capabilities on a months-not-years advancement timeline.
EU AI Office launches first formal investigations under AI Act
The European Commission's AI Office has initiated its first formal probes into potential breaches of the AI Act by frontier-model providers. The investigations focus on obligations around training data transparency, systemic-risk assessments, and red-teaming, establishing reference cases for enforcement across the single market.
Commission issues first detailed guidance on systemic-risk model designation
The European Commission has published its first guidance on classifying AI models as posing a 'systemic risk'. It elaborates on the 10^25 FLOPs compute threshold and introduces capability-based criteria, such as autonomous cyber and code-generation skills, granting the AI Office wide discretion to open assessments.
Germany and France formally establish national AI safety institutes
Germany and France have launched their national AI safety institutes to support EU-level oversight. The German institute focuses on technical evaluations of high-risk models, while France's body will advise sectoral regulators and conduct independent capability assessments, serving as potential templates for other member states.
Benchmarks reveal frontier models' rapid gains in autonomous coding and cyber operations
Recent technical evaluations, including from the UK's AI Safety Institute, show frontier AI models achieving sharp performance jumps on complex software engineering and cyber-security tasks. Models can now autonomously chain vulnerabilities and execute multi-step attack plans, informing regulatory scrutiny on capability-based risk thresholds.
EU and US expand AI chip export controls, widening scope beyond China
The EU and United States have coordinated to tighten export controls on advanced AI chips and chipmaking tools. The expanded measures extend licensing requirements to a broader range of accelerators and restrict transfer of advanced packaging technologies, aiming to manage security risks and prevent circumvention.
AI automation cited as driver in new wave of tech and media layoffs across Europe
A new wave of job cuts in technology, media, and business-services firms explicitly cites generative AI automation as a driver. Investment is shifting from routine content production, support, and clerical roles to AI tooling and engineering, increasing displacement risk for older workers and those in non-specialist positions.
European media and authors file expanded copyright lawsuits against OpenAI and Anthropic
European media groups and authors have filed new or amended lawsuits against OpenAI and Anthropic in courts in France, Germany, and the Netherlands. The claims allege copyright violation from training on scraped content and argue that AI Act compliance should include detailed disclosure of training data sources.
UK AI Safety Institute and EU bodies deepen cooperation on frontier model testing
Collaboration between the UK's AI Safety Institute and EU institutions has intensified, involving sharing red-teaming methodologies, co-developing benchmarks, and exploring joint testing of models. This creates an emerging network of public AI safety institutes spanning the UK, Germany, France, and EU-level bodies.
Energy and water demands of AI datacentres trigger local regulatory responses in EU
Growing concern over the resource consumption of AI-focused datacentres is prompting regulatory responses in several EU member states. Authorities are tightening planning rules, introducing reporting obligations, and requiring operators to use recycled water or invest in grid upgrades, targeting the physical footprint of the AI boom.
This week
- EU AI Office launches first systemic-risk probe into a frontier model's autonomous cyber capabilities.
- National AI safety institutes in Germany, France, Spain begin coordinated testing with Brussels.
- Labour inspectors in Italy, Netherlands, Sweden, Portugal issue first fines for workplace AI violations.
- TSMC and Intel warn that European energy constraints threaten AI chip manufacturing plans.
- Publishers expand copyright suits to demand court-appointed audits of model training data.