Cate shared her story as part of MECEP’s State of Working Maine 2026 report. Click here to read the full report.
Cate is a senior director at a Maine-based human resources technology firm.
In the realm of human resources, we’re presenting AI tools as being there to empower the recruiter. If you interview 10 people for a job, it’s hard to remember them all. It’s there to give you a summary and a transcript. It can also analyze all your previous hires and point to the person that matches your three most successful people who had the job before.
But in reality, we all know that nobody’s building in an audit process, and it’s going to get way too easy to just click a recommendation button. Large companies like Amazon warehouses are already doing first-round interviews with AI. Some companies even have sentiment analyzers that note how often you smile.
You can see how it’s presented as a good thing, and you can also see how it can be a very dangerous thing. For example, at my company, 900 candidates applied to our last open job in a single week. What’s my recruiter supposed to do? Is it better for them to only look at a small number of them, or go until he finds five good candidates and then stop? If you use AI, you can review all of them, and everybody would get the equivalent of a consistent but potentially ragingly prejudiced review by an out-of-control algorithm.
The only way you would know that an AI agent isn’t using discriminatory algorithms is if you’re manually reviewing them at intervals and checking. But that’s not usually happening. If you talk to anyone who’s using AI at scale, it’s like, ok, what’s your audit process? Are you checking 1 out of 100 manually? Do you have a set of records? Do you have a set of benchmarks you consider accurate that you re-evaluate? These things are all being skipped. Because right now, the only way to go at the mythical speed we’re being sold is to not do the double-checking and due diligence and process setting and all these things you actually need to do.
At places where people are working by quota, some managers no longer have the ability to override the algorithm. The more lower level and repeatable a job, those are the ones at risk of being controlled by algorithms.
I am most concerned about CEOs and corporate greed. You can use AI responsibly. It’s the pressure to do it rapidly and produce these magical results that is coming from the nice late-stage capitalism people. That’s what’s making it unsafe. It’s a tool with a lot of opportunity for misuse, and right now the misuse is being encouraged.
The tech giants are getting billions of dollars of investment right now, and they’re only putting out millions in return. The problem with making these models is that it doesn’t get cheaper with scale. It’s not like the first customer cost me $1,000, but the 100th customer only cost me a penny. The first customer cost me $1,000 and the 100th customer cost me $1,000, too. There is no economy of scale. If AI had a killer use case, we wouldn’t be arguing about all of this. There just isn’t one. Nothing people pay lots of money for is made by AI.
I hope that the future of human resources looks like people realizing that finding the right job candidate is a skill that requires a human. Hopefully everyone remembers they should use it as a tool and not outsource their thinking to it. But it will be hit or miss. AI isn’t going away. If people start figuring out now how to take it seriously and use it responsibly and decide to learn this because it’s coming, that is the aspect that is in everyone’s control.