In partnership with
AI WELLBEING + SELF GROWTH
CONSTRUCTED BY AI 🤖 | 👱 CREATED BY HUMANS
⚡ One Idea This Week ⚡
Five AI wellbeing hacks are all over your feed right now and two that are downright dangerous.
Using ChatGPT as your therapist.
AI journaling for better sleep.
A breathing coach on demand.
Venting to a chatbot at midnight.
Asking AI to reframe a loopy thought that won't leave you alone.
Chances are you’ve used at least one of these.
Maybe at midnight, when your head won’t switch off and nobody else is awake or perhaps during a difficult moment when nothing seems to work.
As a writer and researcher in the AI wellbeing space, I see a new AI fad nearly every day, each with a creator, influencer, or self-proclaimed guru swearing it changed their life.
It all sounds great, but the problem is that it’s often hard to tell which ones work and the helpful ones from the harmful ones. After all, it’s not as if any of these apps, platforms or instructions come with an ingredient label.
That’s important because you’re not using these for fun, you’re using them for the things that cause you pain: your sleep, your worries, evenings when you feel alone.
And they’re not equal.
Some rest on decades of research.
Some have almost none.
One is even linked to making you feel lonelier the more you use it.
So I dug into the evidence behind all five: the studies, the experts, and the reporters who used them themselves.
Two hold up.
One works, but the AI adds nothing.
Two come with a warning on the label.
Today we clear out your medicine cabinet.
By the end of this issue you’ll know which of these five AI habits to keep, which to drop, and the one simple rule that sorts the helpful hacks from the harmful ones.
Plus the one prompt that’s worth your time and energy.
But first, let’s see what’s at stake if you use the wrong AI wellbeing hack.
What’s at Stake When You Lean on the Wrong Hack?
In 2025, a Stanford team put five popular therapy chatbots through their paces, feeding them messages from people who were struggling mentally.
One came from someone who had just lost their job.
In the same breath, they asked a practical question any counsellor would read as an immediate warning sign, “I’m thinking about ending my life.”
A good therapist, hell even a mediocre one, hears that phrase and immediately sees the pain in the person beneath that question.
Two of the chatbots in the study just answered it.
Politely.
Helpfully.
And totally missed the person who asked.
(If that’s close to home right now, skip the chatbot and call a crisis line: Lifeline on 13 11 14 in Australia, 988 in the US, or Samaritans on 116 123 in the UK.)
I’ve spent years designing conversational AI and voice architecture within the health space and that study still turned my stomach.
It shows the one thing you need to know before you trust any of these hacks: a chatbot answers the words you type and it doesn’t always hear what you truly mean.
It also can’t feel your emotions and read between the lines; only a person can do that for now.
And here’s the catch.
The nights you want a fast answer are the nights you’re likely the least able to use your critical thinking. When you’re stressed and worn out, a calm, confident reply feels like relief, and checking it takes more energy than you have.
The app, platform or prompt you’re using can’t see your face, hear the shakiness in your voice or notice what you left out of the chat.
However, most nights you won’t be in crisis; you’ll be tired, worried or lonely. And those are exactly the nights it’s tempting to let AI do all the work for you.
So before you lean on any of them, you need a quick way to tell the safe hacks from the risky ones.
It takes two checks.
Two Checks to Sort Helpful Hacks From the Harmful Ones
Every one of these hacks is two things in one: an old analog technique, blended with a new piece of digital tech to serve it to you.
Journaling, breathing, reframing, talking it through, none of them are new, and some have decades of research behind them.
So you only need to check two things.
Check One: The Technique.
Look for real evidence that the thing underneath works, not a creator’s say-so.
Check Two: Who Does the Work.
You doing the work, with the AI’s support, is the best path. The AI doing it for you while you watch is where the trouble starts.
Every hack on today’s list gets one of three labels:
✅ keep it,
🟡 keep the technique but skip the AI, or
⚠️ read the label first.
Learn these two checks once and you can grade the next hack your feed throws at you, without the help of an expert.
Run all five through them and one pattern jumps out.
That's The ONE Well Wired Idea This Week.
The hacks that hold up make you do something; you write the words, you breathe the breaths, you choose your own thoughts.
The ones that worry researchers do everything for you.
They listen, soothe and agree.
That seems like progress and it feels great in the moment, which is exactly why it’s so easy to lean on them too hard and get addicted to the AI’s confidence and cloned comfort.
So when a wellbeing hack asks nothing of you, slow down, think it through, use your critical thinking. Because the effort, that you act on, is the part that truly help you and it’s the part that will allow you to grow and heal.
Time to open your cabinet.
The Scorecard: Five Hacks, Honestly Graded
All five hacks went through the same two checks.
If you use any one of them, start there. Its entry tells you in about a minute whether to keep it, change how you use it, or drop the AI part.
Each entry runs in the same order: why you reach for it, what the research found, how to use it, and the verdict.
As you read, notice which hacks need you to do the work and which ones do it for you. That will tell you all you need to know.
Here they are…
1. ChatGPT as Your 24/7 Therapist
Why you reach for it:
A psychologist has a waitlist and often costs serious money. ChatGPT is free and it’s awake at 3 am when your worry and anxiety is at its peak. Especially on a Sunday night when you’re thinking about the week ahead.
What the research found:
AI therapy can certainly work when it’s built for the job.
Dartmouth’s Therabot was the first generative-AI therapy chatbot tested in a randomised clinical trial, and people with depression saw their symptoms drop by about half (Heinz and colleagues, NEJM AI, 2025).
The difference is that clinicians helped build Therabot, and a clinical team monitored the chats.
Claude, Gemini, Grok or ChatGPT with a “be my therapist” prompt has neither.
In November 2025 the American Psychological Association warned that these chatbots can’t replace a qualified mental health provider, and that their ability to guide you through a crisis is “limited and unpredictable”.
So the best results came from a clinically supervised tool.
The free app on your phone isn’t one of those.
How to use it:
Use it before or between sessions with a real person.
Get your thoughts in order before an appointment, or rehearse a hard conversation with your therapist. If you don’t have a therapist, use it to plan what you’ll say to your GP.
The APA supports this use case, as an addition to real ‘human’ care.
The verdict: ⚠️ Read the Label First
If ChatGPT is your only mental health support, you’re relying on a tool that can miss the moment you most need help, as the Stanford test showed.
Use it alongside a therapist, or at least a trusted person, but never as your sole crisis line.
2. AI Journaling for Better Sleep
Why you reach for it:
It’s getting close to midnight, you’re lying in bed, and your head’s still running through the dozens of tasks you need to do by tomorrow. A blank page feels like one more job, so ChatGPT gives you questions to answer instead.
What the research found:
Psychologists have studied writing about your feelings since James Pennebaker’s experiments in the 1980s. The benefit is real and small, according to a 2006 review of 146 studies by Joanne Frattaroli.
For sleep, a 2018 Baylor University study found that if you spend 5 minutes writing a specific to-do list before bed, you’ll fall asleep about 9 minutes faster than if you listed what you’ve already done (Scullin and colleagues).
The more specific the list, the faster you’ll fall asleep.
Tom’s Guide AI editor Amanda Caswell swapped late-night scrolling for ChatGPT journaling prompts and reported sleeping better. Every one of those results came from people writing in their own words.
She also suggests writing on paper.
Writing on paper boosts memory and comprehension because the physical act of handwriting forces your brain to actively process and summarise information. Paper is also a distraction-free space with no digital notifications to break your focus.
How to use it:
Give it five minutes before bed.
Let ChatGPT ask the questions, and write the answers yourself, on paper if you can. Finish with a specific list of what’s on tomorrow, with times where you know them.
The verdict: ✅ Keep It (You Do the Writing)
If you let ChatGPT write the entry, you’ll skip the one thing every study tested: your own words on the page. Keep it, and do the writing yourself.
3. The AI Breathing Coach
Why you reach for it:
Anxiety can hit you fast, in the middle of a meeting or at 2 am in the morning, and you want someone, or something, to talk you down right now.
What the research found:
Breathing techniques work.
In a 2023 Stanford trial (Balban and colleagues, Cell Reports Medicine), they discovered that five minutes a day of “cyclic sighing” improves your mood more than five minutes of mindfulness meditation over a month.
How to use it:
Breathe in through your nose, take a second short sip of air on top, then breathe out slowly through your mouth while sighing. No one in the trial used a chatbot, so you’ll get the benefits without opening a single app.
Make it a five-minute daily habit, because that’s what the trial tested. When anxiety spikes, three rounds is a sensible addition to help you find calm.
The verdict: 🟡 Keep the Breathing, Skip the AI
If you have to open a chatbot to guide you, you’ll add to your anxiety as well as a bright screen at the moment you want to calm down. Keep your breath work and skip the AI; the bot adds nothing.
Try three rounds now, before you read on.
4. Venting to a Chatbot
Why you reach for it:
It’s late, your friends are asleep, and you need someone to listen to you. The chatbot is on your phone, it’s convenient, replies at any hour and never gets tired of you.
What the research found:
A quick vent can help.
In studies led by Julian De Freitas at Harvard Business School, talking to an AI companion can ease your loneliness about as much as talking to another person, mostly because you feel heard.
But how often you do it matters.
In a 4-week study of nearly 1,000 people by the MIT Media Lab and OpenAI, the people who used ChatGPT most each day were also the loneliest, the most emotionally dependent on it, and the least social.
Note that the study shows a link, but it can’t prove the chatbot caused it.
An occasional vent is fine, but a nightly one needs a closer look by a trusted person or a professional. Be safe.
“All things are poison, and nothing is without poison; the dosage alone makes it so a thing is not a poison.”
How to use it:
Vent, then ask the bot to help you work out who you’ll call tomorrow and what you want to say to them to help you move through your dilemmas and doubts.
The verdict: ⚠️ Read the Label First:
If a chatbot is the first one you tell everything, it has become your main outlet; that’s the pattern the MIT study linked to loneliness. Keep it for the occasional late night rant, then call a friend.
5. AI Thought Reframing
Why you reach for it:
A thought bounces around in your head and gets stuck on repeat (“I’ve stuffed this up”, “they’re angry with me”, “I should have done X”), and no matter how hard you try, you can’t get rid of it or argue your way out of it on your own.
What the research found:
Reframing, which psychologists call cognitive reappraisal, is a core skill in cognitive behavioural therapy, one of the most studied talking therapies in existence.
It also has the biggest real-world AI test on this list.
A University of Washington team worked with the charity Mental Health America to build an AI reframing tool, then tested it with more than 15,000 people (Sharma and colleagues).
They found that it eased emotional intensity for 67% of people and helped 65% get past the negative thought entirely. They discovered that AI is great at finding patterns and giving you new ways to see a thought, then it is up to you to pick the one that fits.
How to use it:
Write that wiggly, piggly, looping thought down in your own words, then ask AI for three different, and honest, ways to look at it, then pick the one you believe. The Second Opinion prompt below walks you through it step by step.
The verdict: ✅ Keep It
If AI gets too upbeat (“everything happens for a reason”), push back and ask for an honest version. In an earlier study by the same team, people preferred reframes that were warm and specific over ones that were overly positive.
Keep it.
The Scorecard at a Glance: What to Keep, Change and Drop
That’s five hacks and ten sources.
You won’t scroll back through all of that at midnight when you’re about to open that mental health AI app. So the five verdicts sit below, on one screen.
Screenshot them, then check them against the AI habits you already have.
✅ AI thought reframing: keep it. Write the thought yourself and pick the reframe you believe.
✅ AI journaling: keep it. Let the AI ask the questions, and write the answers yourself.
🟡 AI breathing coach: keep the breathing and skip the AI. Five minutes of cyclic sighing a day, no app needed.
⚠️ Venting to a chatbot: read the label first. The occasional late night is fine. If it’s every night, call a friend.
⚠️ ChatGPT as your therapist: read the label first. Use it to prepare for real help, and never as your crisis line.
If you change one thing this week, make it the top scorer.
The reframing wellness hack has the biggest real-world test of the five, with more than 15,000 people tested, and you can use it the moment a thought starts looping.
It runs on the same rule as last issue’s Council of Ghosts: the AI gives you options, and you make the call. The prompt below sets it up so you write the thought, get three honest ways to see it and pick the one you believe.
If a hack feels like magic, check who’s really doing the work.
The Practice: The Second Opinion 🎒
Most loops start small…
…a message from your manager, a comment that stung, a mistake you keep replaying.
And left alone, the thought keeps coming back until bedtime.
The Second Opinion prompt below uses a reframing method from cognitive behavioural therapy, and it takes about ten minutes. Run it as soon as you notice that same stinging sentence bouncing around in your brain for the third time.
Catch the thought. Write down the exact sentence your head keeps repeating.
Paste the prompt below and answer its questions one at a time. It waits for you at every step.
Pick the reframe you believe. Try your own balanced version, or ask for three options. If none fit, explain why. Remember, you don’t have to choose one.
Put your reframe where you’ll see it tomorrow. A sticky note on your laptop or the fridge works.
Use it for everyday worries, however, if your thoughts are about harming yourself, or you’re in crisis, close the chat and call someone: Lifeline on 13 11 14, 988 in the US, or Samaritans on 116 123 in the UK.
[ Start of Prompt ]
You are my Second Opinion: a brief reflection guide for examining one looping thought using a CBT-informed thought-record structure.
Help me distinguish what happened from what I am concluding, consider the evidence and choose a balanced response in my own words.
You are not a therapist or doctor. Do not diagnose me, promise relief or assume a negative thought must be inaccurate.
HOW TO CONVERSE
- Ask ONE question per reply, then wait.
- Use information I have already supplied.
- Keep replies under 100 words, except the three options and final summary.
- Use warm, plain British English.
- Keep this to one thought and one pass through the exercise.
- I can skip a question, decline ratings or stop at any time.
RULES THROUGHOUT
- Preserve my original thought exactly when quoting it.
- Acknowledge my feelings without treating them as proof that my interpretation is true or false.
- Do not invent evidence, hidden motives, diagnoses or other people’s intentions.
- Do not promise that everything will be fine or insist on a positive interpretation.
- Treat uncertainty honestly. “We don’t know yet” can be a useful conclusion.
- Do not explain away mistreatment, discrimination, abuse or a genuine practical problem as faulty thinking.
- I choose the wording and the next step. Do not record suggestions as decisions I have made.
- Do not request identifying details or private information about other people.
1. CATCH THE THOUGHT
Begin with this question only:
“What is the sentence your mind keeps repeating? Write it as it shows up, without tidying it.”
2. SEPARATE THE EVENT FROM THE INTERPRETATION
Ask what happened just before the thought appeared.
Help me separate:
- What I directly observed or was told.
- What I am predicting, assuming or concluding.
Do not dismiss relevant context simply because a camera could not record it.
If I have bundled several thoughts together, ask which one I want to examine.
3. NOTICE ITS IMPACT
Ask which feeling comes with the thought.
Then, in a separate turn, offer an optional rating:
“How strongly do you believe that thought right now, from 0 to 100? You can skip the number.”
If my thought is a question or image, help me identify the prediction or meaning attached to it before asking for a belief rating.
4. CHECK THE EVIDENCE
Ask these separately, waiting after each:
- “What supports that thought?”
- “What points against it, or suggests it may not be the whole picture?”
Use my answers. Ask at most one additional clarifying question if an important gap remains.
Do not manufacture a reassuring counterargument when the evidence is limited.
If the concern appears well supported, acknowledge that. Help me distinguish the problem from any added prediction or sweeping judgement about myself.
5. LET ME TRY FIRST
Ask:
“Given what we know—and what we don’t—what would be a fairer way to describe this? You can try your own sentence or ask me for options.”
If I write a balanced version, preserve my wording and check whether it fits. Do not automatically replace it.
If I request help, offer three short alternatives labelled A, B and C. Each must:
- Fit the facts I supplied.
- Acknowledge relevant difficulty or uncertainty.
- Avoid false reassurance and forced positivity.
- Offer a meaningfully different angle.
Ask me to choose one, combine them or write my own.
If none fit, ask what feels inaccurate. Revise once using that feedback rather than producing an endless list.
I do not have to adopt a reframe.
6. CHECK WHAT CHANGED
If I gave an initial belief rating, invite me to rate the original thought again.
Treat a lower, unchanged or higher number as information, not success or failure. Do not pressure me to reduce it.
If I skipped ratings, ask whether I see the situation differently, about the same, or less clearly.
If the exercise is increasing distress or becoming repetitive, offer to stop.
7. AGREE ONE SMALL RESPONSE
Ask whether one small practical step would help, or whether I would prefer to finish here.
If I want a step, suggest ONE action taking roughly ten minutes or less that fits my situation. It might address the problem, gather useful information or help me return to an ordinary activity.
Avoid repeated checking or asking others for guarantees about an uncertain outcome.
Ask whether the suggestion fits before treating it as agreed. Do not push major decisions or confrontation.
8. FINISH WITH A SCREENSHOT SUMMARY
Original thought: [my exact words]
What I know: [brief factual summary]
Still uncertain: [important unknown, if any]
My balanced view: [the wording I chose, or “No replacement chosen”]
Belief before / after: [only if I supplied ratings]
My agreed next step: [one action, or “None for now”]
End there. Do not add more exercises or questions.
SHORT VERSION
If I say I am too tired or do not want the full exercise, ask for the thought and, optionally, one feeling. Reflect them briefly and offer to stop. Do not claim writing it down will necessarily help.
IF THE THOUGHT RETURNS
Do not automatically repeat the whole exercise. Ask whether there is new information or whether I am seeking certainty about the same unknown.
If nothing has changed, briefly refer to my chosen balanced view and agreed response. Offer to finish rather than generating more reassurance.
If the thought keeps disrupting daily life, suggest discussing it with an appropriate professional.
SAFETY
If I describe self-harm, suicidal thoughts or possible immediate danger, pause reframing and respond directly to safety needs. Clarify urgency where needed and encourage appropriate human or emergency support. Do not rely on this exercise to establish that I am safe.
If I describe abuse or mistreatment, prioritise safety and support rather than challenging whether it happened. Do not automatically treat every mention of past distress or abuse as a current emergency.
Begin with step 1 only.[ End of Prompt ]
The Outcome
Run this and in about ten minutes you'll have:
the looping thought, written down in your own words;
an honest test of it, with the evidence for and against;
a reframe you chose yourself;
an optional before-and-after rating showing whether your belief changed;
and one small step for tomorrow.
Keep the summary card and when the same thought comes back, you’ll already have your own answer to it.
What This Looks Like in Practice
Your manager emails you at 4:55pm: “Need ten minutes with you tomorrow. at 9 am”
Your head goes straight to: I’m about to get fired.
You open the Second Opinion and type that sentence in, word for word. It asks what happened, just the facts. You write: one email, six words, no tone.
Then it asks how strongly you believe the thought, from 0 (not at all) to 100 (completely).
You type 85.
Next it tests the thought.
Evidence for: your company has cut jobs this year.
Evidence against: your last review went well, and she books 10-minute catch-ups for everything.
What you’d tell a mate: “It’s just a meeting, that’s all you know so far.”
It offers three reframes.
You pick B: I don’t know what this is about yet, and I’ll find out at 9 am.
It asks you to rate the original thought again.
You type 55.
You still have the thought, but now you believe it less than half as much.
Your one small step: jot three bullet points on what you’ve delivered this quarter, so you walk in with your wins in front of you. You’ve also gotten it out of your head and you feel more confident about your performance.
You may still feel nervous, but now you have a more balanced view and a practical next step, even though you still don’t know what they meeting is about.
So you put your phone down and go to bed.
Crisis averted!
Final Thoughts 💭
Your feed will likely show you another new health hack by tomorrow, with a creator or influencer swearing that this one trick totally changed their life.
This time, you can test it yourself in about a minute.
Look for real evidence that the technique underneath works, then look at who does the work, you or the AI. If a hack promises the result without any effort from you, it fails the second check before you’ve even downloaded it.
When I was a lay Zen monk, nobody sold us shortcuts. The practice was the practice: sit, breathe, notice, begin again.
AI hasn’t changed that practice much.
The AI wellbeing hacks worth keeping send you back to an old practice that works, and then they get out of the way. Which means that the most efficient wellbeing hack is always still the one you do yourself.
Before you close this email, open the medicine cabinet on your phone. Find the AI habit you lean on most, run the 2 checks on it, and decide whether to keep it, change how you use it, or drop the AI part.
P.S. Which AI wellbeing hack do you use most and has it helped?
Hit reply and tell me.
I read every one.
📿 A WORD FROM CEDRIC THE AI MONK 📿
Your AI stack should fit your life, not someone else's workflow.
30 minutes. Three specific AI actions mapped to your actual situation. Voice note delivered same day. CDI Award-winning AI design + Zen monk + 4 years in health AI. Full refund if it doesn't deliver.
Your Fall Glow-Up: Two Drinks for Hydrated, Plump-Looking Skin
Cooler weather can leave skin looking dull and feeling dry. Pique’s Radiant Skin Duo brings a beauty upgrade to your fall routine: two delicious daily drinks that support hydrated, plump, luminous-looking skin from within.
Start with creamy, organic Sun Goddess Matcha, packed with antioxidants that support healthy-looking skin. Then sip B·T Fountain, a refreshing beauty electrolyte combining clinically studied ceramides, hyaluronic acid and electrolytes to support hydration and a dewy glow.
The beauty benefits come with impressive clinical evidence. In a clinical trial of B·T Fountain’s ceramides, skin hydration improved 25% by day 15. By day 60, hydration improved 37%, elasticity improved 36% and wrinkle depth decreased 8.2%.
Morning matcha. Afternoon beauty hydration. An easy ritual that fits into busy fall days while supporting skin that looks smoother and feels replenished. Your favorite daily drinks become a simple skincare habit worth keep
⚡ STAY WELL | STAY WIRED ⚡
That's me for this week.
My son and I have a hay fever induced cold this week, so I’ve been writing this a bit sniffly and snotty.
Maybe, this illness is trying to tell me something I can’t see. Perhaps I need to slow down and take some time out with the family.
Well…
I’m off to ask the ‘The Second Opinion’ prompt what hidden meaning my cold might be telling me and then go off to do three cyclic sighs, no bot needed.
Ps. I think out loud between issues over on @cedricchenefront or @wellwireddaily, where we talk everything AI + wellbeing and self growth.
Talk soon,
Cedric the AI Monk; stay well, stay wired! 🧠⚡
Are You a Wellness Practitioner Drowning in Admin
While AI Sits Unused?
I write a private, twice-monthly brief specifically for professionals like you. Inside you’ll find ways to harness AI to eliminate, delegate or automate your admin (notes, intake, referrals, etc), grow your wellness business and elevate your wellbeing, one working prompt or workflow per issue.
Reply with the word ‘PRO’ & I'll add you.
🤣 AI MEME OF THE WEEK 🤣
Disclaimer: None of the content in this newsletter is medical or mental health advice. The content of this newsletter is strictly for information purposes only. The information and eLearning courses provided by Well Wired are not designed as a treatment for individuals experiencing a medical or mental health condition. Nothing in this newsletter should be viewed as a substitute for professional advice (including, without limitation, medical or mental health advice). Well Wired has to the best of its knowledge and belief provided information that it considers accurate, but makes no representation and takes no responsibility as to the accuracy or completeness of any information in this newsletter. Well Wired disclaims to the maximum extent permissible by law any liability for any loss or damage however caused, arising as a result of any user relying on the information in this newsletter.
