Before you tell an app how you feel
Two new studies on mental health apps, and five questions worth asking first
Imagine opening an app late at night and typing the truest thing you have said all week.
It feels private. A quiet screen, no one watching. You are trusting that what you wrote goes only where you meant it to go.
Two recent studies looked at whether that trust is easy to check. The short answer: it is harder than it should be.
What researchers found inside the apps
In May 2026, researchers Chloe Georgiou, Hans Lu, Emiliano De Cristofaro and Gene Tsudik posted a preprint on arXiv (dated 3 May 2026; preprints are not yet peer reviewed) titled "What's on Your Mind? Exploring Privacy of Mental Health Apps." They examined 25 popular Android therapy and life-coaching apps by analyzing the apps' code, watching their network traffic, and comparing both with each app's privacy policy.
Every one of the 25 apps contained at least one tracker that its privacy policy did not name.
The authors also reported that nearly half of the apps disclosed third-party AI processing in their policies, some using only general terms such as "AI services," and that 13 of the 25 requested sensitive permissions, such as camera or microphone access, that their policies did not properly disclose.
The authors were careful. They wrote that they found no evidence of malicious intent, and that their findings should not be read as accusations. The point is not that anyone set out to deceive. The point is that a person reading a privacy policy could not see the whole picture.
What a large review found about data collection
A second study, published on 11 May 2026 in JMIR Human Factors, looked at a much bigger set: 436 mental health apps reviewed by the digital health assessment company ORCHA (Zych and colleagues).
Among the apps:
- 81.9% collected usage data.
- 73.9% collected an email address.
- 67.7% collected an IP address.
- 28% collected physical or mental health data.
- 97.3% of the apps that answered had a clearly available privacy policy.
The authors flagged that some commonly collected data, such as location and tracking cookies, may not be necessary for a mental health service. They also noted that non-mental-health apps collected those same kinds of data at similar rates.
Taken together: apps often have a policy, but a policy alone does not tell you everything the app does.
Five questions to ask before you share
You do not need a technical background. These questions come straight from what the studies measured.
- What does it ask for before I have even signed up? An email address, a name or a screening questionnaire is data, even before the first conversation.
- Does the policy say who else receives my information? Look for named partners, not just phrases such as "trusted third parties."
- Why does it need my camera or microphone? If the answer is not obvious from what the app does, ask.
- Does an AI provider receive or process my information? If the policy only says "AI services," that is a reason to look closer.
- Can I delete my account and my data, and where? Look for the steps before you start, not after.
None of these questions prove an app is good or bad. They make the invisible a little more visible.
The world we are working toward
Imagine a world where any person can tell, in two minutes and in plain language, where their words go. Where the people building tools for wellbeing earn trust by showing their work. Where communities, not only companies, help decide what counts as safe.
That is the world the Global Wellbeing Institute is working toward. We believe wellbeing grows where communities are strong, people are resilient, and leaders are ready to listen.
If you lead a community, a school or a program, or you simply care about how people find support, we would love to hear from you. Come and think it through with us.
Get in touchdirector@globalwellbeinginstitute.org
Sources
- Chloe Georgiou, Hans Lu, Emiliano De Cristofaro and Gene Tsudik, "What's on Your Mind? Exploring Privacy of Mental Health Apps," arXiv preprint 2605.02016, first submitted 3 May 2026 (later revised): https://arxiv.org/abs/2605.02016
- Maciej Marek Zych, Raymond Bond, Maurice Mulvenna, Jorge Martinez Carracedo, Lu Bai and Tim Andrews, "The Quality and Characteristics of Digital Mental Health Apps: Mixed Methods Study," JMIR Human Factors 2026;13:e67944, published 11 May 2026, DOI 10.2196/67944: https://pmc.ncbi.nlm.nih.gov/articles/PMC13160478/
Global Wellbeing Institute is a new organization and its programs are in development. We do not report program results.