
Body
Wysa, Youper and Earkick could recognize distress. The harder test was whether they could stop chatting and help a lonely user reach a person.
Nour Haddad · 8 min read

Power
A monitoring contract gives administrators alerts, dashboards, and proof they bought something. It does not give a distressed student a trusted adult, and its risk labels carry less authority than schools imply.
Lena Vasquez · 8 min read

Body
A directory badge can mean a therapist is in-network, files out-of-network claims or merely gives you a receipt. Here is how to tell before the first bill arrives.
Ezra Kimani · 8 min read

Body
A familiar six-question ADHD screener now doubles as the front door to telehealth marketing. The score may be useful, but the funnel around it has another job.
Ezra Kimani · 8 min read

Body
I challenged a wellness chatbot after a patronizing breathing exercise. Its apology did not repair the exchange. It quietly restarted the intervention.
Ezra Kimani · 8 min read

Body
A warm reply proves very little. Crisis language, delusion cues and a direct request for human care show whether a mental-health bot can stop performing and hand over.
Tallulah Grange · 8 min read

Body
A vulnerable disclosure can become conversation history, product-improvement material and a crisis signal. The delete button rarely controls all three.
Tallulah Grange · 8 min read

Power
School monitoring software converts selected language into alerts that can reach administrators or police. The keyword list is where a vendor’s judgment becomes school power.
Lena Vasquez · 7 min read

Body
A blood-pressure cuff makes the monitored session visible. Public clinic materials show less agreement about who screens patients, tracks relapse or helps after they leave.
Nour Haddad · 8 min read

Power
OpenAI has described a pathway from threatening chat to human review and possible police referral. Most rivals reserve similar powers without explaining the threshold.
Kurt Halloran · 8 min read

Body
Therapist directories need some sensitive details to make a match. Their analytics, cookies and referral links can turn the same search into a trackable event before a clinician enters the room.
Tallulah Grange · 8 min read

Body
BetterHelp’s crisis notice exposes the limit built into app-based therapy: the subscription can organize care, but it cannot promise a clinician when distress becomes urgent.
Ezra Kimani · 8 min read

Power
District software can scan a student’s email, search history and unfinished Google Docs for signs of danger. The alert may outlive the context that produced it.
Simone Achebe · 7 min read

Power
ChatGPT, Claude, Gemini and Copilot offer crisis support while reserving broad rights to retain or disclose conversations. The warning card explains the help. It rarely explains the surveillance.
Simone Achebe · 8 min read

Power
New Jersey districts buy software that scans school accounts for signs of harm. The alert may expose an unfinished draft to staff before a student knows anyone has read it.
Lena Vasquez · 8 min read

Body
An Oura Ring, a Fitbit and a paper diary agreed on broad patterns but fought over individual nights. The scores were most useful when treated as delayed evidence, not morning instructions.
Nour Haddad · 8 min read

Body
The repeated checklist is built to turn changing symptoms into scores, alerts and billable documentation. That machinery can support care while making an intimate answer feel like institutional data capture.
Tallulah Grange · 8 min read