There was a time when getting help writing a text message meant asking a friend, “Does this sound okay?”
Maybe you were trying to apologize without over-apologizing. Maybe you weren’t sure how to respond to a difficult message. Or maybe you had stared at the same three sentences for 20 minutes, wondering whether adding a period would somehow make you sound angry.
Now, AI can potentially step into that space.
As of August 2026, ChatGPT can connect with Apple’s Messages app on Apple-silicon Macs through an Apple Messages plugin. Depending on how it is configured, ChatGPT can search and read conversations, summarize them, draft messages, and send messages through Messages. By default, users are asked to approve the message and recipient before it is sent.
On the surface, this sounds like another productivity feature. But personal messaging is not just productivity.
Our texts are where friendships happen, relationships develop, families argue, apologies are made, boundaries are set, and sometimes the most vulnerable parts of our lives are shared.
So what happens when AI becomes part of that space?
What can ChatGPT actually do with iMessage?
The new Apple Messages integration goes beyond simply helping you write a text.
On supported Apple-silicon Macs, ChatGPT can search conversations, read message histories, analyze conversations, summarize chats, prepare replies, and send messages through Apple’s Messages app. It can work with iMessage as well as SMS and RCS conversations stored in Messages.
We’ve already become accustomed to asking AI things like:
“Help me write a polite response to this.”
The newer integration moves closer to:
“Look through my conversation with Sarah and tell me what I need to respond to.”
And potentially:
“Draft a response and send it.”
That’s a pretty significant shift. The AI isn’t just helping you compose something you have already decided to say. It can potentially become part of the process of figuring out what to say, how to say it, and whether to say it at all. And that’s where things get interesting from an emotional wellness perspective.
Why might this feature feel appealing?
For some people, messaging can be surprisingly stressful.
A text message can trigger a lot of mental noise:
“Did that sound rude?”
“Should I explain myself more?”
“What if they’re upset with me?”
Having an AI help organize those thoughts can feel like having a neutral sounding board.
It might help turn a long, emotionally charged draft into something clearer. It could help you recognize that a message you’ve written in frustration might be better sent tomorrow. Or it could simply help when you’re tired and don’t have the energy to figure out how to word something.
There can be something genuinely useful about that.
The important question is whether we’re using AI to support our communication or gradually letting it replace our own participation in it.
Could this feature make difficult conversations easier?
Possibly.
Imagine you’ve received a message from someone you care about. You’re upset, but you don’t want to respond impulsively. You could ask ChatGPT to help you identify what you’re actually trying to communicate.
That might create a little distance between the emotion you’re experiencing and the message you’re about to send. Sometimes that may be what we need. Not because our emotions are wrong, but because emotions and communication don’t always move at the same speed.
AI could also help with practical things: making a message less confrontational, making a boundary more direct, or helping someone who struggles to organize their thoughts find clearer language.
But there’s a catch.
A perfectly worded message isn’t necessarily an emotionally honest one.
Could AI-generated messages make our relationships feel less authentic?
This may be one of the more uncomfortable questions. Imagine receiving a heartfelt message from someone you love. They apologize, explain themselves beautifully, acknowledge your feelings, and say exactly the right thing.
Then you find out an AI wrote most of it.
The words might still be meaningful. But you might also wonder:
“Was this actually what they wanted to say?”
That’s where AI-generated communication becomes different from spellcheck or autocorrect. The concern isn’t necessarily that AI-written messages are “fake.” It’s that communication carries more than information. It carries effort, vulnerability, personality, awkwardness, humour, hesitation, and sometimes even imperfect wording.
Those imperfections can be part of what makes communication feel human. A message that takes someone ten minutes to awkwardly write may communicate something different from a perfectly polished paragraph generated in ten seconds. What happens when AI becomes involved in emotionally difficult conversations?
Consider a disagreement with a partner. You could ask ChatGPT:
“Read this conversation and tell me who’s being unreasonable.”
That might feel incredibly tempting.
But now, you’ve introduced AI into a relationship conflict involving two people who have their own histories, emotions, and perspectives. AI can see the messages, but it does not see the relationship.
It doesn’t know what happened three hours earlier. It doesn’t know the tone of someone’s voice. It doesn’t know the inside joke behind a sentence. It doesn’t know what either person is feeling but hasn’t said.
And if you repeatedly ask an AI to interpret your conversations, there is another possibility: you might start outsourcing your own judgment.
Instead of asking yourself:
“What do I think is happening here?”
You may start thinking:
“What does ChatGPT think is happening here?”
That distinction matters.
Could ChatGPT actually help us communicate more thoughtfully?
Yet again, possibly. This is where the technology can be used in a healthier way.
Instead of asking:
“Who’s right?”
You might ask:
“What are the different ways this message could be interpreted?”
Instead of:
“Write a response that makes them understand how wrong they are.”
Try:
“Help me express what I’m feeling without escalating the conversation.”
Instead of:
“Write a response to this message.”
Try:
“Help me think through whether I actually want to write a response to this message right now.”
These are very different uses of AI. Some approaches can turn AI into an emotional referee, while others can use it more like a reflection tool.
What about privacy when ChatGPT can access our messages?
This is one of the biggest considerations.
Apple’s iMessage is designed with end-to-end encryption, meaning message content is protected between the communicating devices. But the new ChatGPT messages integration is a separate layer of access. OpenAI’s release notes say the feature can read and search conversations and prepare / send messages with approval required by default before sending.
In other words, enabling an AI to work with your messages is a crucial privacy decision.
And there’s another factor that may be easy to overlook: Your messages are not only about you. They contain information about other people too.
This may be your partner’s vulnerable text, or your friend’s personal problem, or your family’s health news. These are all real people’s private stories.
Before giving an AI access to your conversations, it is worth thinking about whether you are comfortable bringing those conversations into an AI-assisted environment, not just whether you are comfortable with your own information being accessed.
Could AI change the way we experience emotional connection?
Texting has already changed how we relate to one another. We now communicate through reactions, emojis, typing indicators, read receipts, voice notes, and even GIFs.
AI adds another layer.
Now the words themselves can potentially be generated, refined, interpreted and sent with very little effort from us. This could make communication easier. But easier is not always better.
Sometimes relationships grow through the effort of figuring out what to say. Sometimes an imperfect apology matters because someone struggled to find the words. Sometimes a vulnerable message feels meaningful because the person took the risk of writing it themselves.
If AI removes all this friction from communication, we may also remove some of the moments in which we learn how to communicate.
So, should you use this ChatGPT plugin to help with your messages?
There’s probably no universal answer. Using AI to help organize your thoughts is not inherently unhealthy. Neither is asking for help making a message clearer.
The more important question is what exactly is the role you are giving AI?
A useful rule of thumb might be:
“Use AI to help you communicate, not to avoid communicating.”
Let it help you slow down. Let it help you find the words when you’re struggling. Let it point out how a message might be interpreted
But don’t feel like every emotional conversation needs an AI-generated script.
You don’t have to optimize every text. You don’t always have to find the perfect response.
A Note on AI, Mental Health, and Emotional Well-Being:
Technology can be a useful tool for reflection and communication, but it is not meant to be a replacement for meaningful conversations with the people in our lives or for professional support when we need it.
When it comes to relationships and emotional well-being, sometimes the most helpful next step is not finding the perfect words. Sometimes it is simply having the conversation.
Looking for a therapist?
We have designed a therapist matching quiz to make it easier to find a therapist that fits your needs.
Take the Quiz
Continue Reading
Helpful Resources
Here are a few resources that you might find helpful for your mental health and wellness journey.