BLAST RADIUS

Enterprise readiness · Governance security

Scale Reports Enterprise AI Red-Team Results

Sep 17, 2026

Scale reported that automated graders broke an enterprise AI deployment in 3% of 980 attempts, compared with 68% of human testing sessions.

Read the original at scale.comOpens the publisher's site in a new tab

More in Enterprise readiness