The Future Of Work: Why AI Can’t Write Job Descriptions (Yet)

At least once in each press interview or podcast, I’m asked how I think AI is going to “transform recruiting.” I can’t help but laugh. The Future of Work has been on every event agenda since I started working in this industry in 2009. I have listened to every analysts, sponsors, and “influencer” talk about this topic, each usually offering a slightly different flavor of the same recommendations. Bottom line? AI can’t write job descriptions.

Honestly, I think it’s a little silly. That’s because nothing has transformed in our industry over the last 20 years. Has there been a modernization effort? Yes. There were a lot of companies operating off one spreadsheet on a shared HR desktop computer that now have an ATS. We were still faxing our resume 20 years ago. But a true transformation in recruiting? I think not.

To call it a transformation would imply that something fundamentally changed about how we do the job. That hasn’t happened, even with all the new AI. We still rely on similar systems to get applicants, interview them, and hire the right person even if they have modernized a bit.

Relationships Matter Most

The reason for a lack of transformation is a lack of relationships. Many recruiters are treated like order takers. They’re given the candidate “order” that includes a laundry list of slightly (sometimes, outrageously) unrealistic, made up requirements and are expected to deliver. If they miss one thing in the order, they get criticized and ostracized for not delivering exactly what someone asked for.

Face it. None of the “hot new tech” to make recruiting deliver candidates faster is going to transform that broken relationship. Most tools in this space are focused on speed, not quality, and that will only serve to further deepen the divide between talent acquisition and everyone else. Letting AI write an inaccurate job description quickly isn’t going to make a manager trust you.

If we really want to transform recruiting, it’ll be the consequence of building better relationships and using great tech. The next release of ChatGPT won’t do that by itself.

We Picked The Wrong Things To Automate

But that has always been the message of the future of work – less people, faster delivery. This indoctrination shows up in surveys regarding the top things recruiters want to automate. The top 5 areas I see in every survey of things recruiters want to automate: 

  1. Pre-Screening
  2. Email Communication
  3. Sourcing
  4. Candidate Feedback
  5. Job Descriptions 

Seriously, scan these results. Every survey says the same thing. Here’s the problem. You can’t automate a broken system. And across the board? These are often the most broken parts of recruiting.

PS: I wrote an entire presentation on how you can automate these steps in a way that makes the right humans apply. Book me to speak at your event here.

AI Can’t Write Job Descriptions

Job descriptions are the perfect example. I’ve asked recruiters around the world to tell me if they’re proud of their job descriptions. I’ve seen fewer than 5 hands go up in the last 5 years. They know their job descriptions are awful. They just don’t know where to start on making them better. Naturally, this is the moment where we might turn to AI.

Therein lies the problem. Doesn’t matter how fast you write it if the right person doesn’t read the content and say “I want this job.” If you don’t know what makes a good job posting and you’re not sure of what the hiring manager is really looking for? You can’t write a job post that will deliver the 1 result you need: getting the right person to apply. Let me show you one of these tools in action.

Product Review: Grammarly’s Job Description Generator

The job description is also where that order taker relationship starts. Recruiters create some job writing software or ATS process where hiring managers enter the info, recruiters edit, and post. Now, because there was no hiring manager intake, the hiring manager is treating this like a Domino’s delivery. “I said I want pepperoni. That pizza better have pepperoni.” Or, in candidate terms, “I said 17 years of experience. 16 simply will not do.”

The Future Of Work: One Part Tech, One Part Talent

If we want to transform the job post writing process, it means training your team on how to conduct a hiring manager intake, how to write mandatory requirements based on that conversation (yes, we offer that training), then using AI to get more eyes on the content or create ad variations. AI can’t write job descriptions and it isn’t getting better by letting hiring managers enter their wish list via some recommendation engine that just swaps words that “sound better.” We’re still far away from the place where we can use tech to support and provide options at scale.

That applies to every transformation ahead for recruiting. If we really want change, it’ll take one part tech, one part human. Change will require an upgrade in recruiter skills as humans, not AI operators. If you really want to transform recruiting like all those future of work experts suggest? Change how you work with hiring managers.


Kat Kibben View All →

Kat Kibben [they/them] is a keynote speaker, writing expert, and LGBTQIA+ advocate who teaches hiring teams how to write inclusive job postings that will get the right person to apply faster.

Before founding Three Ears Media, Katrina was a CMO, Technical Copywriter, and Managing Editor for leading companies like Monster,, and Randstad Worldwide. With 15+ years of recruitment marketing and training experience, Katrina knows how to turn talented recruiting teams into talented writers who write for people, not about work.

Today, Katrina is frequently featured as an HR and recruiting expert in publications like The New York Times, Chicago Tribune, and Forbes. They’ve been named to numerous lists, including LinkedIn’s Top Voices in Job Search & Careers. When not speaking, writing, or training, you’ll find Katrina traveling the country in their van or spending some much needed downtime with the dogs that inspired the name Three Ears Media.

%d bloggers like this: