Why I’ll Never Let AI Write My Emails

Need help writing emails? Thanks to the wonders of artificial intelligence and large language models (LLM), you can now have messages composed on your behalf in Gmail, Apple Mail , Outlook, and many other email clients. In most cases, the AI ​​feature appears immediately, ready to provide the assistance you need.

The point is, you can offload routine email processing to AI and move on to other tasks that may be more interesting and important. From personal experience, I’ve spoken with many people who are currently using AI chatbots in exactly this way. But I’m not ready for that, and I don’t think I ever will be.

Here are my arguments, which you may or may not like, although I haven’t mentioned the energy consumption and copyright infringement issues that plague AI use overall. You can keep asking if I need help with Gmail, Gemini, but I’d rather just disconnect you altogether.

I don’t want to forget how to write.

Writing is actually quite easy—most of us can do it from a young age without much trouble. Writing well is more difficult, but you don’t have to be a best-selling author to send a few letters. So is there any harm in using AI to write simple letters? It might be faster and more convenient, but I’m not sure it will actually be beneficial.

As writer David McCullough once said , “Writing is thinking.” Knowing the right word before the next forces the brain to think and process what’s being said. Word choice and sentence structure matter, even in the shortest and most banal emails.

I don’t want to sit down at my laptop one day and find myself struggling to write a few lines of text. Does that seem impossible? Perhaps not, judging by the feedback from those who’ve already tried to delegate email writing to artificial intelligence. “Is it hard to translate that thought from your head into an email?” Google asks . Yes, it is hard, and that’s the point.

People deserve a human response.

It’s safe to say that many of us receive a huge number of emails (if you don’t, consider yourself lucky). Most of those incoming messages are likely from people you don’t know personally, but regardless of the sender or recipient, I believe a human response is worth the effort.

If all our emails—for meetings with colleagues, job applications, project discussions—were written by artificial intelligence, we’d be faced with mountains of machine-generated messages, devoid of any subtlety or personal touch. Imagine a group email conversation where every reply sounds the same, regardless of who sent it.

Even if I’m writing a simple rejection letter, if I’m communicating with another person, I believe they deserve a response from me directly. This is more of a principled position, but I will stick to it.

Artificial intelligence writes a lot of meaningless nonsense.

For example, if you ask AI to write a thank-you note to the organizer of an event you recently attended, you’ll get a rather formulaic text, an amalgamation of countless other thank-you notes. It will be bland, impersonal, and quickly forgotten.

I understand the temptation to use AI to compose an important email—say, when applying for a job or appealing a company decision—but your message will likely read like an algorithmically processed, mass-produced text. Essentially, you’ll sound like everyone else (see the previous point).

I could, of course, have the AI ​​create a draft and then edit it, but if I go that route, I’ll probably end up editing my messages less and less, either out of laziness or habit.

I don’t trust AI to be accurate in detail.

Artificial intelligence still makes plenty of mistakes , though chatbot developers tend to keep a low profile. Whether you’re composing an email about a new project presentation, a family meeting, a client request, or anything else, there’s no guarantee the AI ​​will accurately capture all the details.

The more important the content of an email, the more important the text itself becomes. Companies promoting AI-powered emails seem to believe that we can all use AI to find leads, organize colleagues, and express genuine feelings via email, but I’m not convinced.

People make mistakes too, but I prefer to trust myself rather than the black box of algorithms that even the developers who create them don’t fully understand. Does the AI ​​know who I’m emailing and what details they need? Of course not.

AI interacting with other AI is not the future I want.

To paraphrase George Orwell , if you want to imagine the future, imagine your AI sending thousands of emails per minute to everyone else’s AI, and it goes on forever. At what point will we completely hand over responsibility to chatbots and just let them do their job? I don’t even want to take a step in that direction.

Even the most ardent AI enthusiasts wouldn’t suggest sending AI-written emails without pre-checking and editing, but isn’t that the obvious next step? I can almost picture the Google I/O presentation—outsourcing all of this to Gemini to maximize productivity.

Preliminary research already shows that we forget almost everything we write when using AI, which has alarming implications if we’re sending important information that we’ll need to recall later. I’m not going to buy into that future, no matter how persistent the AI ​​prompts become.

More…