AI use and human oversight

What AI may support

• Research synthesis and working drafts.
• Visual exploration and production assistance.
• Content, code, data, and workflow operations.
• Quality-control checklists and repetitive delivery tasks.

What humans retain

Human team members set strategy, choose directions, review outputs, check product and brand truth, communicate with clients, and approve client-facing work. AI output is treated as material to evaluate, not as an independent decision.

Confidential information

Project inputs are handled according to the applicable agreement and the needs of the work. We do not intentionally submit confidential client material to public AI tools without an agreed reason and appropriate handling. Tool selection may vary by project, so specific restrictions should be confirmed before work begins.

Training and reuse

Round 6 does not use private client material to train its own reusable systems without written permission. Published work, reusable methods, and generalized learnings are handled according to the project agreement and any publication approvals.

Accuracy and product truth

AI-assisted work can introduce errors, invented details, or inconsistent outputs. We review work against available references and intended use. Technical, legal, medical, financial, compliance-critical, or exact packshot requirements may require specialist review or a different production method.

Ownership and disclosure

Ownership, usage rights, source files, licensing, and any project-specific disclosure requirements are defined in the applicable agreement. Clients should tell us about internal procurement, security, legal, or disclosure requirements before production starts.

Questions

To discuss a project’s AI boundaries or data requirements, email hello@round6.co.