Compliance
High-risk content
AI-generated content requiring enhanced safeguards (e.g., deepfakes).
Definition
Generative AI outputs whose misuse or inaccuracy could cause significant harm—examples include deepfakes, sensitive financial documents, or other content affecting safety, privacy, or public trust—where the Guideline recommends additional measures such as tamper‑proof identifiers, traceability, and independent evaluation.
Related Terms
high-risk/high-impact applications
Uses that could materially harm health, safety, rights or environment....
high-risk/high-impact
Uses that could materially harm rights, safety, or environment....
Content
Visual, audio, audiovisual and text-based outputs....
High‑risk AI
AI uses with potential for severe or large‑scale harm....
outputs
Generated artefacts delivered to users (text, images, audio, video)....