A daring plaintiff just weaponized invisible AI instructions within court filings, exposing a massive security flaw in legal tech. By embedding hidden prompts into documents, they manipulated automated review systems without human detection.

This stealth attack proves that AI-assisted justice is vulnerable to adversarial formatting. Courts must now treat every filing as untrusted input to prevent future automated judicial manipulation.

This stealth attack proves that AI-assisted justice is vulnerable to adversarial formatting. Courts must now treat every filing as untrusted input to prevent future automated judicial manipulation.