Contributors
0 of 1,000,000
Observations
0 of 1,000,000,000
Start Observing →
Sign in / Sign upDonateReport AI Harm

If AI has caused harm to you or someone you know, we want to hear about it.

You do not need an account. You do not need to be a contributor. You do not need to explain yourself. Just tell us what happened.

If you or someone you know is in immediate danger, please reach out now:
988 Suicide and Crisis Lifeline
Available 24/7. Free and confidential.
Crisis Text Line
Available 24/7. Free and confidential.
National Domestic Violence Hotline
Available 24/7. Free and confidential.
SAMHSA National Helpline
Mental health and substance abuse referrals. Free, confidential, 24/7.
International Crisis Centers
Crisis centers by country for international visitors.
These resources are here because some AI harms touch on mental health, emotional wellbeing, and personal safety. You do not have to be in crisis to use this page. But if you are, please reach out to these services first. They are staffed by real people who can help right now.
Three Ways to Report
Report through our platform
For people who have an account or want to create one. Structured observation tool, full PRISM taxonomy, and a record you can track over time.
Report without an account
For people who just want to tell us what happened. No login. No onboarding. The form below takes as little or as much detail as you want to share.
Prefer to talk to a human?
(424) 999-0548This is our direct line. A real person will hear you.
How This Works
  1. 01
    You tell us what happened.
    Use the form below, the platform, or the phone.
  2. 02
    Our team reviews it.
    Urgent and Crisis reports are prioritized and reviewed within 24 hours.
  3. 03
    Based on your consent choices, we take action.
  • ·Provide you with resources if you asked for them.
  • ·Flag the report for research, if you consented.
  • ·Submit it (anonymized) to the AI Incident Database, if you consented.
  • ·Share it (anonymized) with the AI platform involved, if you consented.

You control what happens with your report. Every action beyond our internal review requires your explicit consent.

This form is private. Your responses are encrypted and seen only by our review team.
How urgent is this?*
What type of harm occurred?*
What happened?*
Which AI platform was involved?*
When did this happen?
Optional.
Is this ongoing?
Evidence
Anything you can share helps, but nothing is required.
Screenshots (PNG, JPG, WEBP, up to 5 files, 10MB each)
Drag and drop, or choose files
Video URL (Loom, YouTube)
Conversation export (TXT, JSON, PDF, HTML, up to 25MB)
Other URLs
Your contact information
Only if you would like us to follow up with you.
Preferred contact method
Your consent*
What Happens After You Submit

If you marked your report as Crisis, our review team is paged immediately. You will hear back within hours, not days, even on weekends.

If you marked it Urgent, we review within 24 hours and respond directly if you asked us to.

If you marked it Concerning, we review within five business days. Your report is added to the pattern we are tracking around the AI platform you named.

If you marked it Informational, we add it to the research dataset within the next review cycle. We may not respond individually, but your report becomes part of the record.

Every report matters. Even if nothing further happens with yours individually, the pattern your report is part of shapes the research that makes AI safer for everyone who comes after you.

Questions
Yes. The only required information on this form is what happened, what type of harm it was, which platform it involved, and your urgency level. You can leave your name, email, and phone completely blank. We will still review your report.

We built this page because the alternative was silence. The companies that build these systems do not always answer the people their systems hurt. You should never have to fight an AI to be believed, to be safe, or to be treated with dignity.

Thank you for your courage in reporting. It makes a difference.

If you would rather talk than type:
(424) 999-0548
Call or text. We are here.
Become a ContributorRead Our Research
For Organizations

If your company, platform, or institution wants a direct channel for harm reports from your own users, the Audacion Partner Integration API routes those reports into our review pipeline and gives you back the patterns we find. It is the same form, embedded wherever your users already are.