AI in HR: What we're not talking about

David Hanrahan's recent LinkedIn post on the future of AI in HR tech really got me thinking. He raised some key concerns around AI-driven sourcing, job leveling, skill quantification, and performance monitoring that I think are worth diving into further.

AI isn’t taking over HR departments. Not yet, anyway. But it is changing how HR works in subtle, important ways. Take Unilever, for example. They use AI to screen entry-level employees, reporting increased diversity and 75% faster hiring. But questions remain: Is the AI measuring what truly matters? What about candidates who don’t excel at game-like assessments?

Companies are also using AI to predict employee departures and suggest promotions. These systems process data quickly, but speed can compromise accuracy and fairness. Consider Amazon’s “Time Off Task” system in warehouses, which tracks the time between a worker’s package scans. Long gaps can trigger management conversations or disciplinary action.

This approach seems objective at first glance. Everyone faces the same standards. But problems arise. First, it assumes that everything important can be measured. But can you put a number on teamwork? On creativity? On the ability to handle a crisis? Second, it can lead to gaming the system. If employees know they’re being measured on specific metrics, they’ll focus on those at the expense of everything else.

Most importantly, this numbers-first approach forgets that employees are people, not machines. People have good days and bad days. They have lives outside of work. A purely data-driven approach misses all of that. And some companies are going even further, using AI not just to measure performance, but to actively manage people.

Imagine getting your work assignments from an algorithm. Or having your vacation requests approved or denied by AI. It’s already happening in some places. The pitch for this is efficiency because AI doesn’t play favourites. It doesn’t have moods. It can make decisions 24/7. But there are downsides. Big ones.

AI lacks context. It doesn’t know that you’re going through a divorce, or that you just finished a huge project and need a break. It just sees the numbers. It can’t mentor. A good manager doesn’t just assign tasks. They help you grow. They give advice. They support you when you’re struggling. AI can’t do that. Lastly, it can create a sense of being constantly watched. Imagine knowing that every email, every message, every minute of your workday is being analysed. That’s not a recipe for a healthy work environment.

Despite the hype, there’s a lot that AI can’t do in HR. At least, not well. It can’t handle complex human situations. If two team members are in conflict, AI won’t be able to sit them down and work it out. It can’t inspire a team. It can’t give a rousing speech before a big project launch.

And it can’t make the tough, nuanced decisions that HR often has to make. Should we fire a top performer who’s also a bully? How do we handle a sensitive case of workplace discrimination? These situations need human judgment, empathy, and understanding of complex social dynamics.

Most importantly, AI can’t build real relationships. And relationships are at the heart of good HR. This all sounds pretty grim but it doesn’t have to be. Used right, AI could actually make work more human, not less.

How? By handling the routine stuff so humans can focus on what matters. Imagine an AI that could handle all the paperwork for a new hire, freeing up HR to focus on making them feel welcome. Or one that could suggest potential mentors for employees based on their skills and career goals. Or identify teams that might be struggling and need support.

IBM uses an AI platform called “Your Learning” for employee development. It suggests courses based on an employee’s role, skills, and career goals. This enhances, rather than replaces, human development. But consider: How much of your career do you want guided by an algorithm? What about opportunities the AI doesn’t suggest?

AI should support human interaction, not replace it. These tools can help, but they shouldn’t make decisions alone. Why? Because there are serious risks we need to talk about. Privacy is a big one. AI systems in HR often collect a lot of data about employees. Every email, every message, every keystroke could potentially be analysed. How much of your work life do you want an algorithm to see?

Then there’s bias. AI systems learn from data. If that data reflects existing biases — against women, people of colour, older workers, etc. — the AI will amplify those biases. And because it’s AI, it might seem more “objective” than it really is. What happens when the system gets it wrong? If an AI falsely flags someone as a fraud risk, or incorrectly predicts they’re about to quit, that could seriously damage their career. These aren’t hypothetical concerns. They’re happening now, and we need to grapple with them.

AI in HR isn’t just changing how HR departments work. It’s changing work itself. In an AI-augmented workplace, different skills will matter. Creativity, emotional intelligence, and the ability to work well with others will become even more important. Because these are the things AI can’t do.

Company culture will change too. Will employees trust a company where major decisions about their careers are made by algorithms? Will they feel valued as individuals, or just as data points? The answers to these questions will shape the future of work. And right now, it’s up for grabs.

So what should we do about all this? Here are some suggestions:

For HR professionals:

• Learn about AI. You don’t need to become an engineer, but understand the basics.

• Be critical. Question vendors’ claims about their AI systems.

• Put ethics first. Just because you can use AI for something doesn’t mean you should.

For employees:

• Know your rights. Understand what data your company collects and how it’s used.

• Speak up. If you’re concerned about how AI is being used in your workplace, say something.

• Focus on uniquely human skills. Build your creativity, emotional intelligence, and leadership abilities.

For company leaders:

• Think long-term. Don’t just adopt AI because it’s trendy. Consider the long-term impacts on your culture and your people.

• Involve your employees. Be transparent about how you’re using AI and get their input.

• Use AI to augment humans, not replace them. Look for ways AI can help your employees do their jobs better, not do their jobs for them.

Which brings me back to the points David raised. If we’re not careful, AI could lead to managers over-relying on algorithms and neglecting the human side of their roles. It’s crucial that they continue to have genuine, empathetic conversations with their teams.

Similarly, while AI might seem to make things like vacation requests and work assignments more efficient, it risks making employees feel like their individual circumstances aren’t being considered. There needs to be room for human understanding and flexibility.

Fundamentally, AI can’t replace human judgment in HR. It might be able to crunch numbers faster, but it can’t navigate the complexities of interpersonal dynamics, make nuanced decisions, or build real relationships. Those things require human skills and human empathy.

If we can strike this balance, AI has the potential to be a valuable asset in HR. But if we rush to automate everything without considering the human impact, we risk losing the very essence of what makes HR so important — the human touch.

AI in HR is coming, whether we’re ready or not. But we have a choice in how we use it. Will we use it to reduce people to numbers? Or will we use it to help people reach their full potential? That’s the conversation we need to be having. And we need to have it now.

MP