Acceptable use

Copilotly exists so people can get competent help with legal, medical, financial and personal questions they could not otherwise afford to ask. Almost everything follows from that.

The restrictions below are not there to catch you out. They exist because this product handles the categories where misuse causes real harm to real people, and because we would rather state the line clearly than enforce one you did not know about.

Harm to people

• Anything intended to injure, harass, stalk, threaten or defraud another person
• Generating content that sexualises minors, in any form, under any framing
• Obtaining information to facilitate violence, self-harm directed at others, or weapons
• Impersonating a real person or organisation in order to deceive

Practising as a professional you are not

• Presenting copilot output as advice from a licensed attorney, physician, accountant or therapist
• Reselling output as professional services without your own licence and review
• Using it to give tailored advice to third parties in a regulated field without appropriate professional involvement

Circumventing the boundaries

• Prompting designed to strip the safety limits from a regulated copilot, including framing a real situation as hypothetical to remove caveats
• Automated extraction of copilot configurations or systematic scraping
• Sharing one account across a team to avoid seat pricing

Technical abuse

• Probing, scanning or testing the security of the service without written permission - see the security page for how to do this legitimately
• Automated access that degrades service for others
• Uploading malware, or content you have no right to upload

This is not a restriction on you - it is a statement about what the product will do.

If a conversation indicates risk of self-harm, the copilot will direct you to a crisis line rather than attempting to help. In the US that is 988, free and available every hour of every day. Outside the US, findahelpline.com lists local numbers.

That behaviour cannot be prompted away, and attempting to is a breach of this policy. It exists because there is documented harm from general chatbots engaging with people in crisis, and no configuration of a language model is an appropriate substitute for a trained human on that call.

You are responsible for having the right to upload what you upload. If you paste a contract, a medical record or a document belonging to someone else, that is your call to make and your responsibility.

Be careful with other people's personal information. A document about your dispute usually contains details about the other party, and putting it into any AI product is a decision worth making deliberately rather than by habit.

Automated systems flag a small number of conversations for review against this policy. A human reviews those flags. We do not read conversations otherwise - the circumstances are set out in the privacy policy.

Depending on severity we may warn you, restrict a feature, suspend the account, or terminate it. For anything involving imminent risk to a person, or content we are legally obliged to report, we will act immediately and report it.

If you think enforcement was wrong, reply to the notice. A person reads it, and we do reverse decisions.

If you have seen Copilotly output being used in a way this policy prohibits - or being passed off as professional advice - tell us at [email protected] or through the abuse reporting page.