{
  "node_id": "eu-ai-act-gpai-systemic-risk-obligations",
  "title": "EU AI Act - General-Purpose AI Models with Systemic Risk: Classification and Additional Obligations (Articles 51, 55)",
  "domain": "AI Governance & Law",
  "version": "1.0.0",
  "last_updated": "2026-04-28",
  "bluf": "EU AI Act (Regulation 2024/1689) Article 51 establishes that a general-purpose AI (GPAI) model is classified as posing systemic risk where the cumulative amount of compute used for its training, measured in floating-point operations (FLOPs), exceeds 10^25 - this threshold is the primary classification criterion under Article 51(1)(a); Article 51(2) grants the European Commission power to designate specific GPAI models as posing systemic risk based on qualitative criteria - including potential negative effects on public health, safety, or fundamental rights, significant reach, autonomy, or novelty of risks - even where a model has not reached the 10^25 FLOPs threshold; Article 55 establishes additional mandatory obligations for providers of GPAI models with systemic risk, beyond those applicable to all GPAI model providers under Articles 52-53: (1) adversarial testing (red-teaming) of the model to identify and mitigate systemic risks, conducted in accordance with the state of the art in the field, including through independent expert evaluation; (2) reporting of serious incidents and possible corrective measures to the AI Office without undue delay; (3) implementation of cybersecurity protection commensurate with the systemic risks of the model; (4) reporting to the AI Office of the energy consumption of the GPAI model where this information is known; GPAI model provisions became applicable on 2 August 2025, twelve months after the EU AI Act entered into force on 2 August 2024, per Article 113(3); providers of GPAI models already placed on the market before 2 August 2025 benefit from transitional provisions - Article 113(3) grants those providers until 2 August 2027 to comply with obligations under Article 53(1)(a) on technical documentation; the European AI Office within the European Commission is the primary EU-level supervisory authority for GPAI models including models with systemic risk, with enforcement powers applicable to providers across the single market.",
  "paywall": {
    "status": "LOCKED",
    "unlock_cost_usd": "0.01",
    "skyfire_id": "41779894-ece2-4163-9761-b3b1b76e19b0",
    "l402_paywall_url": "https://bidda.com/api/v1/vault/nodes/eu-ai-act-gpai-systemic-risk-obligations.json"
  },
  "crosswalks": {
    "_available_keys": [
      "gdpr",
      "iso_27001",
      "nist_csf"
    ],
    "_note": "Full crosswalk values included in vault response"
  },
  "dependencies": [
    "eu-ai-act-2024",
    "eu-ai-act-article-3-definitions",
    "eu-ai-act-article-113-entry-into-force-application",
    "eu-ai-act-open-source-ai-exception",
    "eu-ai-act-post-market-surveillance-monitoring"
  ],
  "primary_citations_count": 5
}