AI Advisor notice

It is a coach for ethical reasoning. It helps you name the kind of dilemma you are facing, see who has a stake in it, lay out your real options, weigh them against the Field Guide’s frameworks, and rehearse a conversation you are dreading.

• It is not your employer’s policy desk. It has not read your handbook, code of conduct, or contract, and it will not tell you what they say.
• It is not legal advice, and it will not tell you whether something is legal or illegal.
• It does not rule on whether particular conduct broke a rule or a law. That is for the people whose job it is to investigate.
• It does not evaluate, diagnose, or judge named colleagues.
• It is not a reporting channel, a hotline, or a confidential or anonymous service.

If you describe possible misconduct, harassment, discrimination, retaliation, or a risk to someone’s safety, the Advisor will say plainly that this needs a real reporting channel and point you to your organization’s route — and then keep helping you think it through.

Your conversations with the Advisor — the messages you send and the replies you get, with their titles and timestamps — are stored in Ethics Builder’s database (Convex) so the conversation can be there when you come back. A separate counter records how many Advisor messages you sent on each day, and how many tokens those messages used. That counter holds no message text.

• You. Your conversations are yours; no other learner can open them.
• Ethics Builder staff (platform super admins) can access conversations to operate and support the service. We are not a confidential channel and do not claim to be one.
• Your organization’s admins see counts only — how many Advisor messages were sent by their members, and how many of their people used it, per day. They cannot read your messages, your conversation titles, or anything you wrote. There is no screen or export that would give them that.
• Your instructor or manager is not given Advisor data at all.

Anthropic’s Claude, called through Anthropic’s API from Ethics Builder’s servers. Replies are generated by Claude Sonnet; a smaller Claude Haiku model picks which Field Guide chapters are relevant and names your conversation. Under Anthropic’s commercial API terms, what you send is not used to train Anthropic’s public models.

Each person can send up to 40 Advisor messages per day. The count resets at midnight UTC. If your organization has turned the Advisor off, the page will say so.

You can delete any conversation yourself from the Advisor page — open it, then choose Delete from the menu in the top right. That removes the conversation and every message in it. If you would rather not do it yourself, or you want everything deleted, write to hello@meritleadership.com. When an organization ends its agreement, Ethics Builder deletes that organization’s Advisor conversations and their messages.