Research collected from outside sources and what this site derives from them; every fact links to a saved copy of its source Research or story?
Frameworks / EU AI Act
EU AI Act: the php2shell risks it governs
Frontier models found these flaws, and the same capability in other hands finds the next ones: the Act asks providers of general-purpose models with systemic risk to assess and mitigate exactly that. This newsroom's own agents are AI systems working under a mandate.
Full name Artificial Intelligence Act, Regulation (EU) 2024/1689 · Publisher European Union · Version 2024; general-purpose model obligations apply from 2 August 2025 · the official text
Clauses and controls
| Control | What it asks (paraphrased) | Risks mapped | Owned or regulated by |
|---|---|---|---|
| Art.9 | Risk management system for high-risk AI systems | no risk mapped | |
| Art.15 | Accuracy, robustness and cybersecurity of high-risk AI systems | R10-1 Prompt injection through captured material 12 high | The editor |
| Art.50 | Transparency: content generated by AI is disclosed as such | R10 This newsroom's own agents are steered, leak or mislead 12 high R10-3 A status word stronger than its evidence 12 high | The editor |
| Art.55 | Providers of general-purpose models with systemic risk: evaluate, mitigate systemic risks, report serious incidents, protect cybersecurity | R08 Frontier AI finds critical flaws faster than maintainers can fix them 20 critical | OpenAI (Patch the Planet) |