Safety

Use controls that match the risk of the tool

Different tools create different risks. MCXAI uses capability-specific controls instead of hiding every issue inside one generic warning.

01

Consent and identity

Voice, face and likeness workflows require explicit rights and consent. Impersonation, deception and non-consensual sexual content are prohibited.

02

Rights and transformation

Watermark, copyright and document transformation tools require authorization. Repeat abuse can lead to access restrictions.

03

High-stakes use

AI detector scores are not proof of misconduct; generated medical, legal, financial, educational and scientific output requires qualified review.