All resources
Webinars
Transparency Theater: Why AI Disclosure Obligations Increase Your Liability
Say too much and it's evidence. Say too little and it's a violation.
Key Takeaways:
- Delayed isn’t dismissed. EU AI Act and Colorado deadlines moved. Article 50 transparency duties didn’t.
- Deploying AI makes it your problem. Deployers carry their own obligations, even on licensed models.
- Disclosure is a liability paradox. Detail the risks and you’ve documented known harm. Stay vague and you’re noncompliant.
- Every model change breaks your notice. Deprecation, retraining, and new use cases all put disclosures out of sync.
- SaaS contracts won’t cover you. Colorado voids indemnity for a deployer’s own discriminatory acts.
The EU AI Act delay took effect six days before the original deadline, and many teams decided the pressure was off. It isn’t. Every AI disclosure you publish is a live claim anyone can test, and it goes stale the moment your systems change.
Andrew Clearwater, Chief Trust Officer at Airia, and Kyle Miller, Partner at Dentons, show how to keep disclosures defensible without handing regulators a roadmap.
Watch on demand before your disclosure becomes Exhibit A.