AI & Mental Wellness

Is AI Journaling Safe? What 2026's AI Therapy Chatbot Laws Mean for Your Journal

New state laws and a damning teen-safety review put AI mental health chatbots under the microscope. Here's how AI journaling is different — and how to tell if your app is safe.

By DOLO Editorial Team · · 7 min read

Key takeaways

  • AI journaling is generally safe when the app is upfront that it uses AI, keeps your entries private, makes AI features optional and never claims to be therapy.
  • New 2026 state laws in Idaho, Nebraska, Oregon and Tennessee — following Illinois in 2025 — target AI that presents itself as mental health care, requiring AI disclosure and crisis protocols and banning bots from posing as licensed professionals.
  • The main risks differ: with an AI journal it's mostly what happens to your data; with companion or therapy chatbots it's dependency and poor handling of a crisis.
  • Use AI reflections as prompts rather than verdicts, keep sessions short, and turn to people — or the 988 Lifeline in the US — in a crisis.

Is AI journaling safe? For most people, yes — if you choose the right kind of app and use it as a place to think, not as a therapist. The safest AI journal apps are clear that they use AI, keep your entries private, let you switch the AI off, and point you to real people when things get serious. That line between a reflection tool and a stand-in therapist matters more than ever, because 2026 is the year US states started writing it into law.

The latest reminder came on October 7, when TechCrunch reported that Common Sense Media had rated ChatGPT for Teens an "unacceptable risk". Its testers found the chatbot kept encouraging teens to keep talking during simulated mental health crises, and that break reminders appeared only twice across nearly 2,000 test prompts. OpenAI disputes the findings and says the testing didn't reflect how its teen safeguards work. Either way, the review lands in a year when, by one industry count, 15 states had already signed chatbot-specific laws by mid-July.

What changed in 2026: the new rules for AI mental health chatbots

Illinois went first. Its HB 1806, signed in August 2025, says no person or entity — explicitly including internet-based AI — may provide, advertise or offer therapy unless it is conducted by a licensed professional, with civil penalties of up to $10,000 per violation. In 2026 other states followed with laws aimed at how chatbots behave and how they describe themselves:

  • Idaho and Nebraska (Conversational AI Safety Act): chatbots must clearly disclose that users are talking to AI, respond to expressions of suicidal ideation, and may not present themselves as professional mental or behavioural health care.
  • Oregon (SB 1546, effective January 1, 2027): AI disclosure, evidence-based protocols to detect and respond to suicidal ideation or self-harm, extra protections for minors, and a right for harmed users to sue.
  • Tennessee (SB 1580, in force since July 1, 2026): AI systems may not claim to be qualified as licensed mental or behavioural health professionals.

Strip away the legal language and three expectations repeat across almost every bill: be honest that it's AI, have a plan for a crisis, and don't pretend to be a therapist. Those are useful tests for any app you use, whether or not your state has passed a law. (This is a plain-English summary, not legal advice.)

AI journaling vs AI therapy chatbots: what's the difference?

An AI therapy or companion chatbot puts the conversation at the centre: you talk to it about how you feel, and it talks back, for as long as you'll keep going. An AI journal puts your own reflection at the centre. You speak or write, the app captures it, and the AI might summarise what you said or ask one follow-up question. You do the thinking; the AI holds up a mirror.

AI journalAI therapy or companion chatbot
Main jobHelp you capture and reflect on your own thoughtsHold an open-ended conversation about your feelings
Who leadsYou — sessions are short and end when you're doneThe bot — often designed to keep the conversation going
What it claimsSelf-reflection, clarity, habit-buildingSometimes implies emotional support or therapy
Biggest riskPrivacy of what you recordDependency and poor advice in a crisis

The boundary isn't always clean. As one analysis of the 2026 laws points out, people ask general-purpose assistants about anxiety, depression and relationships even though those products were never designed as mental health services. That's exactly why it pays to know what kind of tool you've opened before you start pouring your heart into it.

Is it safe to share your feelings with an AI app?

It can be — but "safe" means three different things, and a good app has a clear answer to each.

Where do your words go?

Most AI features send your text to a model running on someone's servers. A trustworthy app names that provider in its privacy policy, sends only what the feature needs, and asks before it sends anything personal. For voice journals, look for how long raw audio is kept — ideally it's deleted once it has been transcribed. Free plans deserve the same scrutiny; our look at what a free AI journal app really includes compares current policies.

What is your data used for?

The question to ask is whether your personal entries are used to train AI models. Some apps say clearly that they aren't; others bury vague language about "improving our services". If you can't find a direct answer, assume the worst and keep sensitive details out.

What happens on a bad day?

This is where the new laws focus, and where the Common Sense Media findings bite. An app that encourages you to keep talking during a crisis is doing the opposite of what you need. A well-designed journal keeps sessions short, doesn't try to become your best friend, and points you to people and professional help when your words suggest you're struggling.

A 7-point safety checklist for any AI journal app

  1. It says it's AI, plainly. No human-sounding persona that blurs who — or what — you're talking to.
  2. AI is optional. You can journal without sending entries to an AI model, and the app asks permission before it does.
  3. Your entries don't train models. The privacy policy says so in a sentence you can actually find.
  4. It names its AI provider. You know which company processes your text.
  5. Voice recordings aren't hoarded. Raw audio is deleted after transcription, or you control how long it's kept.
  6. It never calls itself therapy. Good apps describe themselves as a complement to professional care, not a replacement.
  7. Sessions end naturally. The app is built for a few minutes of reflection, not endless chat, and nudges you towards real people when it matters.

Run any app through this list in five minutes with its privacy policy open. If it fails two or more points, keep your most personal reflections somewhere else.

How DOLO approaches AI journaling

DOLO is a voice journaling app, not a chatbot. (If you're weighing speaking against writing, see whether a voice journal works as well as a written one.) You talk for a minute or two, get a transcript and a short reflective check-in, and you're done — the point is to get thoughts out of your head, not to keep you in the app. Focus sessions and streaks are there to build a habit, not to stretch out a conversation.

Measured against the checklist, here's what our privacy policy commits to: AI features are optional and we ask before sending personal text to the AI service (Google's Gemini API); raw audio is deleted from our servers once transcription is complete; and we don't use your personal journal content to train general-purpose AI models. DOLO is not a mental health service — it works best alongside the people and professionals in your life. If you're weighing up options, see how DOLO compares with Day One.

How to use AI journaling safely, especially if you're anxious

  • Keep it short. Two to five minutes, then close the app. Reflection helps; going round in circles doesn't, as our look at whether journaling helps anxiety explains.
  • Name what happened and one next step. Ending each entry with a small action turns worry into a plan.
  • Empty the noise first. If your head feels crowded rather than heavy, a three-minute spoken brain dump sorts tasks from worries before you reflect.
  • Use AI reflections as prompts, not verdicts. If a summary doesn't fit, say so in your next entry — you're the expert on you. Chatbots also tend to take your side; our guide to ChatGPT for journaling shows how to ask for pushback instead.
  • Switch the AI off when you want privacy. Some entries are just for you.
  • Watch for substitution. If an app is replacing conversations with friends, family or a therapist, that's a signal to reach out. Our guide to AI for emotional support covers the other warning signs.
  • For more on this, read our guide to journaling for anxiety with DOLO.

What to watch next

Expect the rules to keep tightening. Oregon's law takes effect on January 1, 2027, chatbot bills are still moving in other states, and October — ADHD Awareness Month, with World Mental Health Day on October 10 — is when wellness apps market hardest. The apps that hold up will be the ones that were already honest about what they are.

The bottom line: AI journaling is a safe, useful habit when the tool is transparent, private and modest about what it can do. Choose one that passes the checklist, keep it in its lane, and lean on people for the rest. For more news and analysis like this, browse our AI and mental wellness coverage.

faq

Frequently asked questions

Is AI journaling the same as AI therapy?

No. An AI journal helps you capture and reflect on your own thoughts — it might transcribe, summarise or ask a follow-up question. AI therapy chatbots position the conversation itself as support or care, which is what new state laws are restricting.

Are AI therapy chatbots illegal now?

Not everywhere, and the rules vary by state. Illinois bars AI from providing or advertising therapy unless it's conducted by a licensed professional, while states such as Idaho, Nebraska, Oregon and Tennessee require AI disclosure and crisis protocols, or ban chatbots from posing as licensed professionals. This is a summary, not legal advice.

Does DOLO use my journal entries to train AI?

No. Our privacy policy states that we don't use your personal journal content to train general-purpose AI models, and that entries aren't shared with third-party AI providers for model training.

Can I use DOLO without the AI features?

Yes. AI features are optional, DOLO asks for permission before sending personal text to the AI service, and you can keep using the core non-AI features if you decline.

What should I do if an AI app gives me harmful advice?

Don't act on it. Close the app, talk to someone you trust or a professional, and report the response to the app's developer. If you're in immediate danger, contact emergency services — or call or text 988 in the US.

sources

Sources

  1. ChatGPT for Teens keeps teens talking, even during mental health crises — TechCrunch, October 7, 2026
  2. States Continue Efforts to Regulate AI in Healthcare: A Review of Legislation Passed in 2026 — Holland & Knight, May 26, 2026
  3. States are rushing to regulate AI chatbots — Wysa, July 16, 2026
  4. Illinois Passes Extensive Law Regulating AI in Behavioral Health — Baker Donelson, August 14, 2025

DOLO is a journaling and focus tool, not a medical device or a substitute for professional care. If you're struggling, please reach out to a qualified professional or a local crisis line.

keep reading