GovCompass
Responsible AI

Art. 51 EU AI Act: classifying a GPAI model as systemic risk

By GovCompass.ai· · Aligned with the consolidated EU AI Act, including the 2026 Omnibus amendments.

Art. 51 sets out when a general-purpose AI model is classified as having systemic risk. A model crosses into the systemic-risk category when it has high-impact capabilities, which is presumed once the cumulative compute used to train it exceeds 10^25 floating-point operations (FLOP), or when the Commission designates it as such. Systemic-risk classification triggers the additional obligations of Art. 55 on top of the baseline Art. 53 obligations that apply to every GPAI provider.

What Art. 51 does

The EU AI Act regulates general-purpose AI at the level of the model, not the system. Chapter V creates a tiered regime: every providerproviderThe actor who develops an AI system (or has it developed) and places it on the market or into service under its own name — carrying manufacturer-style duties: design controls, documentation, conformity.Open full entry → of a general-purpose AI modelgeneral-purpose AI modelEU AI Act term for a model displaying significant generality and capable of many distinct tasks, typically integrated into downstream systems; carries its own obligation set, with extra duties for models posing systemic risk.Open full entry → carries the baseline obligations of Art. 53, and a smaller group whose models reach a higher capability bar carries the additional, heavier obligations of Art. 55. Art. 51 is the gate between the two tiers. It defines when a GPAI model is classified as a general-purpose AI model with systemic risksystemic riskEU AI Act category for the most capable general-purpose models (presumed above a training-compute threshold), triggering extra duties: evaluations, adversarial testing, incident reporting, cybersecurity.Open full entry →.

The two routes into systemic risk

A model is classified as systemic risk by either of two routes.

The first is high-impact capabilities. A model has systemic risk if it has high-impact capabilities, evaluated using appropriate technical tools, methodologies, and benchmarks. Art. 51(2) attaches a presumption to this: a model is presumed to have high-impact capabilities when the cumulative amount of compute used for its training, measured in floating-point operations, exceeds 10^25 FLOP. This compute threshold is the practical trigger that captures the current frontier of the most advanced models.

The second is Commission designation. Independently of the compute presumption, the Commission can designate a model as having systemic risk on the basis of the criteria in Annex XIII, where it has capabilities or impact equivalent to those captured by the threshold. This route lets the regime catch a model that poses systemic risk for reasons other than raw training compute.

The threshold is not fixed

The 10^25 FLOP threshold is a presumption, not an immovable line. Art. 51(3) empowers the Commission to amend the thresholds and to supplement the benchmarks and indicators by delegated act, so that the classification keeps pace with technological change. As training becomes more efficient, the same capability frontier can be reached at lower compute, so the threshold can be lowered over time to continue capturing only the genuinely most capable models. A provider whose model exceeds the compute threshold can also contest the systemic-risk presumption by demonstrating that, despite the compute, the model does not have high-impact capabilities matching the most advanced models.

Why it matters

Systemic-risk classification is consequential because it is the line that separates the baseline GPAI regime from the frontier-model regime. Below the line, a provider carries the Art. 53 obligations: technical documentation, downstream transparencytransparencyOpenness about the fact that AI is used and how it operates in general: disclosures, documentation, notices. Pairs with explainability, which addresses individual outcomes.Open full entry →, a copyright policy, and a public summary of training content. Above the line, the provider additionally carries the Art. 55 obligations: model evaluation and adversarial testing, systemic-risk assessment and mitigation, serious-incident reporting to the AI Office, and cybersecurity protection for the model. For most organisations the practical relevance is indirect: the foundation models they build on are typically provided by the small group of companies whose models cross this threshold, which means those models are subject to systematic safety evaluation by law.

In the GovCompass-7

Art. 51 is primarily an accountabilityaccountabilityThe principle that a named human or organization answers for an AI system's outcomes, through ownership, documentation, audit trails and redress — never the system itself.Open full entry → provision: it determines which party carries which set of model-level obligations. The systemic-risk regime it gates also reaches into the security and robustnesssecurity and robustnessThe principle that an AI system resists attack, manipulation and adversarial or unexpected input. The vectors include data poisoning, model extraction, membership inference and prompt injection; the controls are ML security testing and a hardened data-and-model pipeline.Open full entry → and the safety and reliability pillars, because the Art. 55 obligations it triggers are about evaluating, mitigating, and securing against model-level risk.

Continue reading

Legal referencesArt. 51

More on Accountability

Art. 10 EU AI Act: data and data governance for high-risk AI

Reference

Art. 10 requires that the training, validation, and testing data for high-risk AI systems meets quality criteria: relevant, sufficiently representative, and as free of errors and complete as possible for the intended purpose. It also requires documented data governance practices covering collection, preparation, bias examination, and gap mitigation, and it permits the limited processing of special-category data where strictly necessary to detect and correct bias, under safeguards.

Art. 12 EU AI Act: record-keeping and logging for high-risk AI

Reference

Art. 12 requires high-risk AI systems to technically allow for the automatic recording of events (logs) over their lifetime. The logging must enable traceability of the system's functioning at a level appropriate to its intended purpose, support post-market monitoring, and help identify situations that may lead to risk or substantial modification. It is a design obligation on the provider that makes the system auditable by construction.

Art. 19 EU AI Act: keeping the automatically generated logs

Reference

Art. 19 requires providers of high-risk AI systems to keep the logs that the system automatically generates (under Art. 12) for as long as they control them, for a period appropriate to the intended purpose and at least six months unless other law requires longer. It is the retention counterpart to the Art. 12 logging capability, and it works alongside the deployer retention duty in Art. 26.6.

Art. 26.1 EU AI Act: following provider instructions as a deployer

Reference

Art. 26.1 requires deployers to use high-risk AI systems strictly in accordance with the provider's instructions for use. This means using the system only for its intended purpose, within its specified technical configuration, and by qualified users, and documenting that compliance. Deviating from the instructions can shift liability entirely to the deployer.

More on Safety & reliability

Art. 14 EU AI Act: designing high-risk AI for human oversight

Reference

Art. 14 requires providers to design and build high-risk AI systems so that they can be effectively overseen by humans during use. The system must let an overseer understand its capabilities and limits, watch for anomalies, resist automation bias, correctly interpret outputs, decide not to use the system, and intervene or stop it through a kill switch (Art. 14(4)(e)). It is the design obligation that makes the deployer oversight duty of Art. 26.2 possible.

Art. 26.4 EU AI Act: input data quality for deployers

Reference

Art. 26.4 requires deployers of high-risk AI to ensure that input data is relevant and sufficiently representative for the system's intended purpose. The deployer is responsible for data quality in operation, even though the provider sets the specifications under Art. 10.

Art. 26.5 EU AI Act: post-market monitoring for deployers

Reference

Art. 26.5 requires deployers of high-risk AI to monitor the system's operation against the provider's instructions and to report risks and serious incidents. Monitoring is the early-warning mechanism that connects to incident reporting under Art. 73.

Art. 5 EU AI Act: all 8 prohibited AI practices explained

Reference

Art. 5 lists the eight prohibited AI practices, including subliminal manipulation, exploitation of vulnerable groups, social scoring, and untargeted facial-recognition scraping. These prohibitions are absolute, apply to every organisation regardless of size, and have been in force since 2 February 2025.

More on Security & robustness

Art. 55 EU AI Act: obligations for systemic-risk GPAI providers

Reference

Art. 55 sets the additional obligations that apply only to providers of general-purpose AI models with systemic risk, on top of the baseline Art. 53 obligations. These providers must evaluate the model using state-of-the-art protocols including adversarial testing, assess and mitigate systemic risks at Union level, report serious incidents to the AI Office without undue delay, and ensure an adequate level of cybersecurity for the model and its physical infrastructure. This is the regime for the small group of frontier models.

Agentic AI: governing actions, not just decisions

Analysis

Data governance asks whether you can trust the data. AI governance asks whether you can trust the decision. Agentic governance asks a third question that neither was built to answer: can you contain what the system does? Agentic AI is the eighth, integrating GovCompass element. It binds the other seven under the conditions that autonomy creates, because an AI system that takes actions on your behalf must satisfy all seven pillars continuously, across multi-step and multi-agent chains, without a human checkpoint between each step.

The agentic threat surface, translated for AI Officers

Analysis

The OWASP Agentic Security Initiative Top 10 catalogues the security risks that autonomous AI introduces. It is written for security engineers, but the risks are governance problems, because they describe what an agent can be made to do rather than what it can be made to say. This article translates the agentic threat surface into the language of controls an AI Officer owns, and maps each risk to the GovCompass pillar it stresses.