New AI Tool Flags Moments in Simulator Sessions for Instructors to Review — Not to Score

AviatePro Services and simulator manufacturer Simnest Group have developed an AI-assisted training-review tool called Simvidence that analyses pilot behaviour during simulator sessions — tracking facial movement, voice tone, gaze, head orientation and head-movement speed — to flag moments that deviate from an individual pilot's own established baseline for closer review during debriefing. Crucially, the system is explicitly designed not to score or grade pilot competency itself; AviatePro founder and chief executive Krisztian Makai said plainly that "AI should support the instructor's judgement, not replace it or determine competency grades," positioning Simvidence's output as discussion prompts for instructors rather than automated verdicts. Makai described the tool's function with a similarly modest framing, saying "we don't want to provide mind reading. We're just reading the signs," emphasising that instructors retain full responsibility for actual assessment. Simnest Group, which combines a flight-simulator manufacturing arm with its own pilot academy training roughly 800 cadets annually, is using the tool within its own training operations as both a proving ground and a data source. The companies said their longer-term goal is to collect a million hours of simulator training data to validate and refine what the system flags, an approach the announcement compared to 19th-century physician Carl Wunderlich's methodical body-temperature research establishing what counts as a normal baseline before abnormalities can be meaningfully identified.