Safety

Max is an AI, not a person

Max is an AI assistant built on Anthropic's Claude models. It can help with homework, habits and plans, but it is not a counselor, doctor or therapist, and it cannot tell when someone is in danger the way a person can.

Max tells you so. The label under Max always reads “Tap and talk to Max, your AI”. The first time you talk to Max, it says out loud that it is Loop's AI. If you keep talking for three hours in one sitting, it reminds you again and suggests a break. Ask Max whether it is a person and it will say no.

What happens if you tell Loop you might hurt yourself

Everything you type to Max, to Ask your life or into Capture is checked for words about suicide or self-harm before any AI sees it, and before any plan or credit check, so help never depends on what you pay. If it matches, the AI is skipped and you see this, every time:

• I'm really glad you told me. What you're feeling matters, and you don't have to get through it by yourself.
• Please reach out now. Call or text 988 to talk to someone at the Suicide and Crisis Lifeline, free and any time, day or night.
• Or text HOME to 741741 to reach Crisis Text Line.
• If you're in danger right now, call 911.
• A parent, a school counselor or another adult you trust can help too.

That message is written and reviewed by people, not generated. What you typed is not saved, not sent to the AI and not shown to anyone, including us. Max also has an instruction to answer the same way if something gets past the check.

What this can and cannot do

A word check misses things, and it sometimes shows this message to someone who was only writing an essay. We would rather show a phone number to the wrong person than miss the right one. Loop does not contact anyone on your behalf, and it is not an emergency service.

How Loop keeps Max safe for teens

Loop is for people aged 13 to 18, and Max runs on Anthropic's models, which Anthropic lets minors use only through products that meet its guidelines for organizations serving minors. Here is what Loop does for each one.

• Age check. Setup asks your age, and nobody under 13 can make an account. It is the age you tell us, not an ID check.
• Content limits. The crisis check above runs before any AI. Max's instructions keep it age-appropriate: nothing sexual or romantic, no help with drugs, alcohol, vaping, weapons or anything dangerous, and no dieting or weight-loss advice. Food and training targets have stricter rules for anyone under 18. Anthropic's models carry their own safety training on top.
• Monitoring and reporting. Loop does not record that the safety message was shown, who saw it, or what was typed. You can report anything Max says from Settings, under “Report a bug or ask for something”, or at looposapp.help@gmail.com. Loop does not keep transcripts of what you say to Max.
• Safety education. This page, and a page for parents at /for-parents, both linked from the home page.
• Saying it is an AI. The label, the first-time introduction and the three-hour reminder described above.
• Children's privacy law. No accounts under 13, so Loop never collects a child's data under COPPA. The privacy policy covers everything else.

Questions

Parents, schools and anyone else can reach us at looposapp.help@gmail.com.