Responsible AI

Everything we ship touches AI systems — our own products, and the assistants our clients advertise in. This is the written framework we hold that work to. It is deliberately proportionate to our size: one accountable owner, clear criteria, and a bias toward pausing first.

Ownership

Responsible AI ownership sits with the founder, Yann Borie (Superstellar LLC, Zug). Every question, concern, or incident lands on that desk directly — there is no committee to get lost in. Reach the owner at hello@superstellar.io.

Use-case authorization

New AI use cases — in our products or in client work — require owner approval against written criteria:

• Platform-policy compliance.We operate within the usage and advertising policies of the platforms we build on, including OpenAI’s ad policies, Google Ads and Microsoft Advertising policies.
• No prohibited advertising verticals. We do not run campaigns for gambling, adult content, weapons, political advertising, or unsubstantiated health and financial claims.
• Disclosure by default.Every paid placement is clearly labeled as advertising. No system we build may influence — or be marketed as influencing — an AI assistant’s organic answers. The ad lives alongside the answer, never inside it.
• Data minimization. No data-broker audiences, no re-identification attempts, no targeting on protected categories.
• Human approval for generated content. Customer-facing AI-generated content ships behind a human approval step — ad copy is approved by the advertiser before it serves; the same pattern holds across our products.

Escalation

Policy questions and incidents go directly to the owner. The default action is pause first, investigate second: affected placements or features are stopped before root-causing begins, and platform partners are notified where their policies are implicated.

Review

This framework is reviewed quarterly and whenever a platform we operate on materially changes its policies.

ADOPTED JULY 2026 · OWNER: YANN BORIE · Superstellar LLC, ZUG