Why you might not want to use AI to write your cover letter

There are two promises of generative AI that seem tailor-made for the job search process. One is that it eases repetitive tasks. The second is that it provides valuable suggestions that could improve your writing. On the surface, these two benefits are perfect for a job search. You have to write a ton of cover letters, so anything that speeds the process is a blessing. On top of that, you may not be confident in your writing skills, so getting a generative AI system to punch up your letter a bit is likely to make you come across better to hiring managers.

But . . . in this situation, the benefits of AI may be smaller than you might expect.

The fundamental problem is bound up in a paradox highlighted in a paper in the journal Science Advances by Anil Doshi and Oliver Hauser. They find that when people engage with generative AI to create a short story that it improves the creativity of the stories they generate. But, if you have a group of people who all use generative AI to develop their stories, the collective output of the group is less creative overall than it would be without using generative AI.

At first, this might seem strange. How could everyone get more creative, but the group gets less creative. The answer is that generative AI gets everyone to think a little differently about the project they’re doing. But, engaging with a generative AI model is like having every single writer ask for advice from the same person. Each person is getting similar advice, and so the collective output gets more homogeneous than if every writer got advice from a different expert.

A similar thing happens when you use generative AI to help with your cover letters. You could ask a generative AI model to write a letter based on a prompt you give it. Or you could feed it a draft you wrote and ask for suggestions. Either way, the letter you get back is likely to seem like a good one to fit your purpose. So, you’re likely to feel happy that you got help with the letter.

The problem is, probably lots of other job applicants are doing the same thing. It would basically be as if they were all going to the same person (or perhaps the same very small group of people—given that there are a few different foundation models out there). So, the hiring manager is deluged with letters that all sound pretty similar.

Hiring managers are likely to develop an immunity to AI-generated letters, and will start paying attention to letters that feel more personalized and idiosyncratic. Indeed, in a world of copycats, your brain naturally pays more attention to things that feel novel.

Perhaps an ideal strategy is to play with large language models a bit to get a feel for suggestions they make that could improve your cover letters. These models might suggest a couple of words or phrases that you like. But, after that initial exploration, consider writing you own letters. This helps you develop a natural style that will sound more authentically like you. In the long run, that will help you stand out in a crowd.