WarriorIQ legal · draft for counsel review

AI Transparency Notice

WarriorIQ uses computer vision and temporal models to track selected fighters and propose fight events. Customer reports are automatic, and outputs are labelled when they are preliminary or uncertain.

Policy version 2026-08-24
Draft policy architecture—not legal approval. Final wording, jurisdiction, operator details and effective dates must be reviewed by qualified counsel before public launch. View the launch gate.

1. What the AI does

Local models estimate person boxes, pose, identity continuity, motion and action candidates. Ruleset logic filters legal techniques and an evidence layer decides what may be shown. Optional OpenAI identity recovery sends selected frames only after explicit upload-time opt-in and only when configured.

2. What the AI does not do

It does not provide an official WAKO result, biometric identification of a real-world identity, medical advice, or a decision with legal or similarly significant effects. It can confuse fighters, limbs, techniques, timing, contact or ruleset outcomes.

3. User control

Users choose Fighter A and Fighter B once, then the performance report runs automatically. No scorecard labelling or correction work is required. Users can replay supported evidence, and unsupported measurements remain hidden or clearly unavailable. Preliminary estimates never become verified facts merely because tracking coverage is high.

4. Quality and complaints

Tracking coverage is a system observation metric, not a guarantee of correctness. Report material errors through support and include the analysis identifier rather than sending the original filename in public channels.

5. Contact and operator

Verified operator, jurisdiction, and monitored support details will be published here before paid public launch.