Q: I’ve started receiving emails from companies and organizations that all seem strangely impersonal. Even marketing messages and fundraising appeals sound stiff, generic and oddly robotic. Are these being written by AI, and if so, is there any way to tell?
A: An AI-powered LLM is undoubtedly writing some of those messages, but it’s getting more and more difficult to tell. That’s by design. AI systems are trained on vast quantities of online writing, including mountains of marketing materials, so are remarkably good at mimicking a polished, friendly, slightly enthusiastic tone. Honestly, I’ve read emails written by humans that sound more robotic than the best of those generated by AI.
Dave Taylor / Technology
As these tools become increasingly sophisticated and people learn more about how to use them, it’s going to become impossible to distinguish between a good AI and a mediocre human writer.
But let’s look at this a different way.
Whether an email was written by a person or by ChatGPT, Claude, or another AI might not actually be the important question. Better to evaluate whether the message is useful, accurate and honest.
I use AI to help brainstorm ideas and organize my thoughts, but I still review every word before I publish anything. Most modern businesses are doing something similar: A human comes up with the concept, AI drafts the message, then the human personalizes it before it lands in your inbox.
So can you identify AI writing?
Nonetheless, there are some clues that suggest AI might have had a hand in the writing. AI-generated messages often sound polished but generic. They avoid strong opinions, repeat the same idea in different words, and have a habit of saying things like, “We hope this message finds you well” or “In today’s fast-paced world…” They’re grammatically flawless but feel completely impersonal.
Dry, stilted writing is not proof, however, or we’d have to throw away 90% of school textbooks. Plenty of organizations intentionally write in that style because they’re trying not to offend anyone. AI simply learned by reading millions of examples of that exact style. If anything, it’s become an expert at corporate-speak. Somewhere, there’s a committee discussing whether they’ve been replaced by an AI-powered counterpart.
You can even paste a suspicious message into an AI chatbot and ask whether it thinks another AI wrote it. Yes, there’s some irony in the analysis, but the answer isn’t especially reliable.
That’s why it makes more sense in this AI era for us to stick with the basics; Judge the message, not the author. If someone is asking for money, personal information or login credentials, slow down and verify the request independently. If it’s a newsletter from your favorite charity or local business, the words may well have started life in an AI chatbot before a human polished them up, or vice-versa, started as human prose polished by an AI program. That’s becoming increasingly common, and there’s nothing inherently wrong with it.
After all, most of us don’t care whether a calculator helped prepare our tax return. We care whether the numbers add up. AI-generated writing deserves the same treatment. The real measure isn’t who wrote it. It’s whether it’s worth your time.
Dave Taylor has been involved with the online world since the beginning of the Internet. He runs the popular AskDaveTaylor.com tech Q&A site and invites you to subscribe to his weekly email newsletter at AskDaveTaylor.com/subscribe/ You can also find his entertaining gadget reviews on YouTube at YouTube.com/AskDaveTaylor.