Use AI thoughtfully
AI, Trust & Boundaries
Models can help interpret and produce information. They do not decide their own access, actions or authority inside Big AI Bro.
AI can be useful and still be wrong
AI may misunderstand information, miss context, rely on imperfect sources or produce an incomplete or incorrect result. Confident wording is not proof. Important information should be checked against suitable evidence.
Capability is not permission
A model's technical ability does not expand a Specialist's job. The platform defines what knowledge is relevant, what a Bro may access, which actions are permitted and which layered approvals must pass.
Access is never assumed
Deploying a Bro does not give it access to everything. Access must be explicitly available, purpose-bounded and permitted by the account and any connected system. Access alone does not authorize an action.
Approvals are layered
An action proceeds only when the platform, Specialist, Bro lifecycle, configuration, authenticated account, external permissions and safety rules permit it—and when any specifically required human approval has occurred.
Evidence and history matter
Research-capable work must distinguish evidence from analysis and uncertainty. Applicable execution, artifact and delivery history is preserved so the result can be tied to the Bro, configuration and exact Specialist Version that produced it.
Check consequential decisions
Do not rely solely on AI output for medical, legal, financial, safety or other high-impact decisions. Use appropriate qualified professional advice and verify important facts.
Product boundaries are not a privacy policy
This page explains the product model. Formal privacy and legal commitments remain in their own documents and must be approved before public launch.