Search the Atlas

Search risks, controls, and glossary terms

MediumIP InfringementRealized

IP / copyright infringement

Content Safety & Integrity

Description

Outputs reproduce copyrighted text, code, or trademarked material, creating infringement exposure.

Example scenario

Generated marketing copy closely reproduces a competitor's copyrighted material.

Real-world evidenceRealized

In May 2023, an AI-generated fake image of an explosion near the Pentagon spread on social media and measurably moved the S&P 500 index briefly before being debunked. This is a confirmed production incident where AI-generated content directly affected financial markets, even though it was not created by a financial institution's own system.

Primary mitigations

  • Output similarity/plagiarism checks
  • licensed-source grounding
  • copyright filters
  • indemnified providers.

Detection signals

Plagiarism/overlap detection; copyrighted-snippet matching.

Mitigating controls

4
Non-agentic controls

Related risks in Content Safety & Integrity