Quite often, I’ll hit “tab” and let Gemini respond to an email for me. As someone who gets hundreds of emails a day, I’ve found this to be one of the most effective ways to ensure I respond quickly and efficiently.
But as a writer and a human, it also gives me an icky feeling, knowing that I’m giving minimal effort for the sake of speed. I’m well aware that maybe the recipient of my email used AI to write theirs; we do live in an era where AI email assistants are doing a lot of the heavy lifting.
It’s all to easy to take a look at what the response Gemini has prepared for us in faded grey text, hit tab and send.
That’s why, a month ago, I decided to try it on myself. Instead of having Gemini write emails for me, I wanted to see what would happen if I let it write emails to mebased on my emails to myself. Here’s what happened.
The experiment
Every day, I wrote myself an email. Sometimes it was a recap of everything I’d accomplished. Other times, it was a messy collection of unfinished thoughts, frustrations, ideas and things I didn’t want to forget. I didn’t worry about grammar or structure. Nobody else was going to read it.
It was kind of like I did the work of an AI-wearable such asBee or Looki, and then told AI all about it. Then I sent the email to myself and let Gemini respond. This was somewhat similar to using ChatGPT as a journal, but there was no prompt, just myself talking to myself and then giving it to AI to reply.
To be clear, I wasn’t looking for therapy or a digital best friend. I was curious whether an AI that spends so much time helping people communicate could do anything useful when the person on both ends of the conversation was me.
Here’s what I learned
1. Gemini sounded like an HR department at first
Unlike using Gemini as a chatbot, when it’s connected to your Gmail, it feels oddly professional. In fact, the first thing I noticed was how shallow and strange the response felt. Sure, it was kind enough, but it was just so surface.
Everything was an opportunity. Every frustration was a learning experience. Apparently, even complaining to myself required a positive attitude and three actionable next steps.
Sometimes, that’s all someone needs, especially when we need to keep it professional, but it was definitely a wakeup call to stop using Gemini to write responses for people I actually know.
In another email to Gemini, I had written something about a project that kind of flopped and Gemini responded with the sort of relentlessly upbeat language we’ve all come to recognize from AI-generated emails.
Everything was an opportunity. Every frustration was a learning experience. Apparently, even complaining to myself required a positive attitude and three actionable next steps. And while it was funny at first, it made me think how often we send this kind of language to actual people.
When I read those responses knowing exactly what I’d written and how I’d felt, the disconnect was impossible to miss. Gemini could produce a perfectly acceptable email without really capturing the point of what I was trying to say.
2. It picked up on things I’d almost forgotten
Once I got past the corporate pleasantries, the experiment became more useful, especially as I often used my emails to not just talk about my day but send myself reminders. I’d write “The kids have school spirit week for the next few days, so I have to remember that Wednesday they need to wear blue and on Thursday they can wear pajamas to school. Oh, and Thursday all three kids have dentist appointments…”
I liked using this opportunity to write down my thoughts in the most messy way possible. My daily emails were not organized or planned at all. I’d start with something that happened at work, jump to an idea for another project, remember something I needed to do tomorrow and then circle back to whatever had been bothering me that afternoon. I liked that I could bounce around and Gemini caught everything; it could pull individual details out of that mess and put them into some kind of order.
Sometimes it highlighted a task I’d mentioned almost in passing. Other times, it identified a practical problem buried beneath several paragraphs of complaining.
That was the part I hadn’t expected. I already knew what had happened during my day. I didn’t need AI to tell me. What I needed was a way to look at everything I’d written without getting distracted by the thing that had annoyed me most.
Gemini wasn’t always right about what deserved my attention. But even when I disagreed with its priorities, I found myself thinking more carefully about my own.
3. The most useful response wasn’t always advice
At first, I expected Gemini to give me answers. If I wrote about something frustrating, I wanted a solution. If I mentioned an unfinished project, I expected a plan. But I started getting more out of the experiment when I asked it to do less.
Instead of requesting a list of recommendations, I asked Gemini to identify one thing I might be overlooking or one question worth considering. That made the responses more interesting and much more personal.
Instead of requesting a list of recommendations, I asked Gemini to identify one thing I might be overlooking or one question worth considering.
That made the responses more interesting and much more personal. And while sometimes the question was obvious, other times I realized I had missed the point or needed a different perspective.
Occasionally, it forced me to think about something I’d been avoiding or hadn’t realized was connected to another problem. It also made the routine feel less like asking AI to manage my life and more like using it to organize my own thinking because I realized it was surprisingly useful to write everything down without restrictions.
How to try it yourself
This experiement is really easy and interesting. If you want to try it, all you need to do is make sure Gemini is active in your Gmail.
Start with an email to yourself. Spend 5 or 10 minutes writing about your day. Include what went well, what didn’t, ideas you want to revisit and anything you’re worried about forgetting. Don’t edit yourself.
Ask Gemini to respond. You can use Gemini’s features in Gmail, where available, or paste the email into the Gemini app. I found that giving it specific instructions was more useful than simply asking for a reply.
Here’s a prompt to try:“Read this email as a daily debrief, not a message that needs a professional response. Skip the pleasantries and motivational language. Tell me what I accomplished, identify anything important I may have overlooked and ask me one question that could help me approach tomorrow differently. Keep it conversational and don’t assume every problem needs a solution.”
Keep the emails. Create a label in Gmail so you can find them later. At the end of the week, reread what you wrote and compare it with the responses. You may notice recurring problems, unfinished ideas or progress that was easy to miss from one day to the next.
One caution: avoid including confidential work information, sensitive personal details or anything you wouldn’t be comfortable sharing with an AI service. Check your account’s privacy and data-use settings before trying this.
My verdict
After a month, I don’t think Gemini understands my life better than I do. And I certainly wouldn’t trust it to decide what I should do next. But the experiment changed how I think about AI email assistants.
I’d started with a tool designed to help me communicate with other people. What I found was a way to make better sense of the things I was already telling myself. And overall, the biggest surprise was learning how insincere Gemini’s responses felt to the person (me) on the receiving end.
I plan on continuing the daily emails and leaning on Gemini for perspective, but when it comes to using Gemini to send emails to other people, I think I’m done doing that for now.
- Amazon and DoorDash take photos of your front door — but good luck finding out when they’re deleted
- Meta Muse can keep files on people who have never used it — here’s what I found
- I made 7 Gemini Skills to handle my most repetitive prompts — #4 is my personal favorite
