AI-supported underwriting is now mainstream, but accountability still sits with the underwriter. This panel examines how carriers define the boundary between machine recommendation and human authority. When does automated triage make sense? When must judgment override model output? How are underwriting leaders setting escalation thresholds and validation controls? Speakers will share how underwriting teams are structuring governance around AI-assisted pricing, appetite alignment, and eligibility screening. The discussion focuses squarely on underwriting decision authority, not abstract AI debates, and how leaders maintain discipline while increasing speed.