This as-told-to essay is based on conversations with Matt Stauffer, a 41-year-old CEO of Tighten, a programming company. He is based in Atlanta. His words have been edited for length and clarity.
I’m the CEO of a small programming company. I recently did my first full round of hiring in the age of AI. Reading through applications and narrowing down the pile has always been a time-consuming process. This time, it was significantly harder.
We always try to include some personal, fun questions in an application to get a sense of who someone is. This hiring round, I put it to a higher degree because we’re trying to understand people’s humanity a little more now with AI. How do they communicate, what are they passionate about, and how do they communicate about what they’re passionate about?
We had 703 applications for an open project manager role, and I read all of them. About 30 applications in, I could tell some of them were fully AI. Everybody has a different relationship with AI, and we’re open to people at varying levels of engagement, as long as your brain is still involved and the final output is good.
Having signs of AI writing does not mean you won’t get a job with us, but there is a line.
I tried to make the job application more personal this time
On August 29, I went live with my project manager job posting. We’re a remote company of around 15 full-time team members, and because we’re fully remote, much of our interaction happens over Zoom or in writing. We wanted to suss out who’s capable of doing those things at a professional level.
We can’t talk to over 700 people, but we want to really understand them as people as much as possible, while still trying to honor the time that applicants are taking out of their day. Especially right now, when the hiring world is miserable, and a lot of people are filling out thousands of applications.
Some of the personal questions I included in the application were things like, ” How do you pronounce GIF?” There’s no right answer; it’s just for fun to see how they engage with the question.
Other questions include, “What’s your favorite movie?” “What are your thoughts about the Oxford comma?” and “What is the most formative experience in your journey toward becoming the project manager you are today?”
There were a few dead giveaways for the AI answers
I studied English in school, I write blogs, and I wrote a book. Regardless, if you’re on the internet at all, at this point, you can probably tell what AI writing sounds like.
One of the most obvious ones was the movie question. The AI, I believe, has decided there are a few movies that everybody should list as their favorite: “The Pursuit of Happiness,” “Inception,” “The Secret Life of Pets,” and “The Lion King.”
If someone was using AI to fill out the application, they had also given information to the AI agents about the context in which this question is being asked. So, I read a few applications where it had responses like, “I really love Inception,” and then the applicant went on to write 10 sentences explaining that they like “Inception” because they’re a project manager and project managers like movies like “Inception.” It doesn’t feel like the appropriate tone or communication style for a question about your favorite movie.
We also would see, for example, someone accidentally copy the boilerplate around AI’s answer. So the answer would start with, “Here’s a great answer you should put that will show how great of a project manager you are.”
For the Oxford comma question, the application would always say “reduces ambiguity in lists.” I know the models are non-deterministic, so AI gives a little bit different output every time, but there are certain things it homes in on, and after reading a few applications, I thought, Oh, if it says reduces ambiguity in lists, there’s a really good chance this is an AI answer.
I’m not actually against people using AI in their applications
I use AI primarily in writing and coding, as more of a reviewer and conversation partner. Occasionally, I use it to help avoid staring at a blank screen, even if I know everything that ended up on that screen would be thrown away later. I’m not against AI at all; it’s a great tool. I just get concerned because if I can tell you used AI, clients can too.
I’ll see applications where there are signs that they used AI, but clearly, there’s a human involved, who prompted, reviewed, and edited their own humanity into it. Also, there’s always a chance that I’m wrong. It might actually be fully human, and it’s just a human who happens to write that way, so I need to give space for those people.
AI is a tool, which means it’s not like we should never use it. But like any tool, we should know how to use it well. I want to make sure that my team is capable of writing if Claude goes down or something similar to that happens.
Do you have a similar story to share? If so, you can reach out to the reporter at aapplegate@businessinsider.com
Read the full article here
