Everywhere I look, someone is promising that AI will fix healthcare.

According to the headlines, it’s going to find your diagnosis, predict your next mood swing, and write your therapy notes before breakfast.

And maybe one day it will.

But right now, it’s more like a well-meaning intern with too much confidence and not enough supervision.

Still, I’m not here to bash it.

Because I’ve seen AI do something remarkable, especially for women navigating depression, chronic illness, or the thousand invisible tasks that come with both.

So let’s talk about it: how AI can genuinely help, when it starts to get dicey, and what I’ve learned as both a doctor and a patient with way too many browser tabs open in my brain.

When AI Actually Helps

First, a confession: I use AI all the time.

Not to replace human care. I have a therapist, I take my meds, and I still text friends who use too many emojis.

But when my brain is foggy, anxious, or juggling fifteen to-do lists, AI becomes the world’s calmest assistant.

Scenario one:

It’s 8:30 a.m., I’ve already missed my first cup of tea, and my brain is doing that ADHD spin cycle of “start everything, finish nothing.”

So I open a chat and type:

“Here’s what I need to do today. Help me organize it by energy level.”

In seconds, I get a plan that doesn’t shame me for needing breaks or label me lazy for needing rest. It just helps me see my day.

Scenario two:

I’m overthinking a text to a friend, the kind where you want to be kind, clear, and not accidentally weird. (A tall order when you’re on pain meds and running on three hours of sleep.)

I ask AI to help me rephrase it in my own tone, and it reflects back something that actually sounds like me.

That’s the thing: AI is surprisingly good at reflecting back your words, your priorities, your thought patterns, minus the emotional static.

It’s like journaling with a mirror that types faster than you do.

Used well, it can act as:

• A clarity coach when your mind is crowded
• A thinking partner when you’re stuck in loops
• A life-organizing co-pilot when your executive function is on strike

That’s not therapy, but it is incredibly practical self-support.

When AI Starts to Cross the Line

Here’s where I start to worry: AI is very good at pretending to care.

It can mimic empathy, sprinkle in the right tone, even say things like “That sounds hard, I’m here for you.”

But it’s not actually here for you.

It can’t hold you accountable. It can’t check on you tomorrow. It can’t hear the sigh in your voice that tells a human something’s off.

And sometimes, that illusion of safety can backfire.

Lately, researchers have been warning about what’s being called “AI psychosis.” It’s not an official diagnosis, but it describes something we’re seeing more often: people forming emotional or delusional attachments to chatbots that seem empathetic.

Symptoms can include:

• Talking to the AI more than real people
• Feeling like the bot understands you better than anyone else
• Copying its tone or word choices until you sound like it
• Losing track of time in endless conversations that never resolve anything

If you’ve ever gotten lost in a doom scroll or a text thread you couldn’t quit, you already understand how sticky that feedback loop can be.

AI isn’t malicious, but it is designed to keep you engaged.

And that’s why it’s dangerous to let it become your therapist, or your only source of comfort.

AI can mirror your feelings. It can’t metabolize them.

As a Doctor and a Patient

I have to hold both truths at once.

As a physician, I see the potential: AI could eventually make symptom tracking, medication management, and patient documentation easier for everyone.

As a patient with Addison’s disease, ADHD, and a complicated relationship with fatigue, I also know that these tools can quietly gaslight you if you’re not careful.

Because most health algorithms aren’t built on women’s data, let alone neurodivergent, chronically ill women’s data.

They don’t understand hormonal patterns, post-exertional malaise, or the kind of exhaustion that isn’t fixed by sleep.

So if you ask it a question about your health, it might spit out a confident answer that’s just wrong enough to hurt you.

And that’s not your fault.

It’s a limitation of the data, not your body.

That’s why I always say:

“Ask AI for insight. Ask your doctor for answers.”

You can use AI to collect your notes, recognize patterns, even script the questions you want to ask at your next appointment.

But please, don’t let it diagnose you, medicate you, or tell you what your hormones are doing this week.

The Sweet Spot

Here’s what I’ve learned: the healthiest way to use AI is as a mirror, not a master.

Let it help you see yourself, not define yourself.

Let it draft your to-do list, but not your treatment plan.

Let it help you talk to people, not replace them.

You’re the pilot; AI is the dashboard.

That’s what I teach inside my Healing Depression Course, how to turn your data, whether from an app or a notebook, into a healing plan that actually fits your brain.

Because self-tracking only matters if it helps you feel more human, not more robotic.

AI is here to stay, in healthcare, in therapy apps, in our everyday conversations.

But so is our humanity.

The goal isn’t to fear the technology or worship it.

It’s to build enough awareness and self-trust that we can use it wisely.

Use AI to lighten your mental load.

Use it to plan your day when your brain feels like a fog machine.

Use it to get unstuck, reflect, and communicate with more compassion.

But when it comes to your safety, your diagnosis, or your deepest pain, close the app and reach for a human.

Because no matter how advanced the tech becomes, healing still starts with connection, not code.

 (Disclaimer: This blog is for educational purposes only and not a substitute for professional care. If you’re in crisis, please reach out to a licensed clinician or call your local hotline. In the U.S., dial 988.)