There’s a really interesting new trend in the world of recruiting: deciding between technology and people to solve problems. I’m sure that doesn’t come as much of a surprise to you. If you subscribe to any of the same recruiting newsletters I do, you’ve read the same headlines about “eliminating recruiters” and the “future of recruiting.”
These headlines are surprising to me. I’ve been writing, researching, and studying HR and recruiting for the last 15 years. There have been a million tools, platforms, and everything in between introduced to this audience to remove bias. But tools have traditionally been marketed as an option to enhance the desired outcomes – move faster, prioritize, etc. Never have I heard tools mentioned as an alternative to people.
Realistically, I find myself laughing when technologists say AI is replacing people now. As recruiters, we’re dealing with one of the most unpredictable variables in the world. How could an AI tool automate a process most can’t explain and isn’t the same for each person? News flash: you can’t (yet).
Actually, AI Can’t Replace People (Yet)
I believe this perception that AI can replace people starts at the fact that for the first time, technology is generating answers instead of options. In simple ways, AI mirrors the interaction with an expert on the team minus those pesky follow-up questions people like to ask (note sarcasm).
Studies confirm that tools are often more influential than a person would be, down to the way the people talk to each other. For example, a study showed that if AI referred to a symbol in a specific way, people on the team started to shift their language in real time, referring to the symbol the same way as AI. People on the team changed their language not because they knew AI was telling the truth. Not because they were confident it was correct, but just because the AI called it that first. It’s the digital equivalent of wearing Adidas in the 7th grade. All the cool kids were doing it, so you did too.
The reality? AI isn’t always right. It’s not always telling the truth. Which is exactly why it’s so important to have training for teams – not just on how to use technology, but the part humans do best. The part where we actually guide hiring managers to determine what is true, what our goals are, and how to tackle bias in real time. We need people who know how to mitigate the bias of the machine.
Three Ways To Use AI To Remove Bias From Recruiting Messages
The reality here is that all AI tools have bias because everyone (and all the data) used to build them have bias. It’s as simple as that. In the future of work, we need both humans and AI if we want to drive more ethical outcomes in the world of hiring. But first, you have to train recruiters how to do this.
Start with what content. One of the simplest ways to remove bias from your content is by auditing the formatting. Here are 3 ways to do that.
- Bullet Overload: When we have more than 7 bullets or 1/3 of the page is a bulleted list, you will see fewer responses from female candidates whether it’s a job posting or an email. There are parallel studies for underrepresented groups that suggest the longer the list, the fewer candidates from diverse backgrounds you’ll see.
- Gendered Language: While this influence has been negated in some ways, it’s still important to understand that hyper-masculine “bro” language won’t entice most candidates to apply.
- Preferred Requirements: They are not universally understood. Some people read preferred and understand it’s nice to have. Others will opt out – especially people from traditionally undervalued groups. Don’t assume anyone understands. Add context around preferred requirements. For example. “If you meet any of the following, you qualify for a higher base salary.”
Once you train your team to look for those biases, you can use ChatGPT to enhance your content well beyond some gender decoder. Here are some of my favorite prompts for removing bias.
Recruiter Training Must-Have: How To Remove AI-Generated Bias
It’s less important to learn how to use tools right now and more important to learn how to vet them for bias, the limits of technology, and how to validate outcomes. These are lessons vendors won’t teach during tech onboarding when you spent a million bucks and no one wants you asking those pesky human questions.
I’m offering customizable training for teams on how to remove bias from their AI generated hiring content. Whether you’re just getting started or you have some experts on the team, there’s something for everyone in this presentation. I include the key areas of bias in automation, ethical concerns to keep in mind, and build in practical application to improve how your team is using AI now. Let’s get everyone to the same level of understanding about what AI is (and isn’t), how to make their job easier with AI, and the inevitable bias.
Let me train your team on the questions to ask, red flags to look for, and get everyone feeling more confident we’re building a better future of work – not just an AI-powered one. Book a meeting here to discuss bringing it to your team.