1. Verify Clinical Endorsements & Clinical Foundation: Is the app based on established psychological theories or evidence-based practices?

• Ask:
  • Is the app endorsed by medical institutions (e.g., APA, NHS)?
  • Does it cite peer-reviewed studies proving efficacy?
• Red Flags:
  • Vague claims like “AI-powered therapy” without evidence.
  • No mention of evidence-based practices (e.g., CBT, mindfulness) or endorsement from reputable professionals or institutions.
• Example: Woebot, an APA-reviewed app, openly shares clinical trial results showing reduced anxiety in 70% of users.

---

2. Scrutinize Data Privacy Policies

• Ask:
  • Is data encrypted? Who owns it?
  • Is user data sold to third parties or used for ads?
• Green Flags:
  • Apps like Ash AI Therapy store data locally (not on servers).
  • Clear GDPR/HIPAA compliance statements.
• Red Flags:
  • Buried privacy terms allowing data sharing.

---

3. Check for Clear Disclaimers

• Ask:
  • Does the app state it’s not a substitute for human therapy?
  • Are crisis protocols provided (e.g., suicide hotline links)?
• Example: Replika includes disclaimers like, “I’m an AI companion, not a licensed therapist.”

---

4. Assess User Feedback & Safety

• Ask:
  • Do reviews mention harmful advice or dependency?
  • Is there a way to report unsafe interactions?
• Red Flags:
  • Apps with lawsuits (e.g., Character.AI’s teen suicide cases).
  • No moderation of AI responses.

---

5. Look for Human Oversight

• Ask:
  • Can users escalate to licensed professionals?
  • Does the app partner with clinics (e.g., Talkspace integration)?
• Green Flags:
  • Youper offers live therapist sessions.
  • Calm connects users to crisis counselors.

---

Your Quick-Start Guide: 5 Questions to Ask Before Trying an AI Therapy App

Use this checklist as a first step—not a substitute for professional advice—to spot red flags in AI mental health tools.

Question	Notes
1. Does the app have clinical endorsements or peer-reviewed studies?	□ Yes □ No □ Unsure
2. Is your data encrypted, and does the privacy policy ban third-party sharing?	□ Yes □ No □ Vague
3. Does it clearly state it’s NOT a replacement for human therapy?	□ Prominent □ Hidden □ None
4. Are crisis resources (e.g., hotlines) provided for emergencies?	□ Yes □ No □ Hard to find
5. Can you escalate to a human therapist if needed?	□ Yes □ No □ Paid add-on

Next Steps:

• For deeper vetting: Use the APA’s App Evaluation Model to assess clinical relevance and safety.
• When in doubt: Consult a licensed therapist before relying on AI tools.

---

Why This Isn’t Enough

This checklist is designed to filter out blatantly unsafe apps—not guarantee effectiveness. Even apps that pass these checks may:

• Lack cultural sensitivity for non-Western users.
• Use biased algorithms (e.g., underdiagnosing marginalized groups).
• Fail during technical glitches, leaving users stranded.

Always pair AI tools with human oversight.

---

Pitfalls to Avoid

• Over-Reliance: AI can’t diagnose disorders like depression.
• Hidden Data Sharing: Free apps often monetize your mental health data.
• Bias: Apps may perform poorly for non-Western or disabled users (APA, 2023).

---

For Deeper Evaluation: APA’s App Advisor

The American Psychiatric Association’s App Evaluation Model offers a step-by-step framework:

1. Basics: Privacy, security, transparency.
2. Evidence: Clinical validation, research backing.
3. Usability: Accessibility, engagement.
4. Clinical Relevance: Alignment with treatment goals.

---

Disclaimer

This article is not medical advice. AI therapy apps are supplemental tools and should never replace licensed professionals. If you’re in crisis, contact a therapist or emergency services.