Prompt injection exposes a wider AI governance gap: what GCs should do now
September 04, 2026
Prompt injection exposes a wider AI governance gap: what GCs should do nowSeptember 04, 2026 Lexology Pro In this article examining how concealed instructions in litigation documents could manipulate AI tools, partner Amy Rudolph discusses practical safeguards for externally sourced documents. As Ms. Rudolph notes, this risk is not limited to litigation documents but can extend to any process in which businesses rely on AI to analyze documents received from third parties. She outlines steps legal teams can take to reduce risk, including screening for warning signs, cleaning documents before upload, instructing AI tools regarding embedded instructions, and independently verifying important outputs. Key contacts
Latest Events |