When Isaac Asimov wrote his three laws of robotics in the 1940s, he was writing science fiction. But the problems he was trying to solve — how do you govern a powerful, non-human intelligence so it serves people rather than harms them — turned out to be more foresight than fiction.
Most of us know those laws from the Will Smith movie I, Robot: a robot may not harm a human being, a robot must obey orders given by humans, and a robot must protect itself. They were Asimov’s attempt to build guardrails into something that, without them, could go badly wrong.
The talent world been thinking a lot about those laws lately, because as AI works its way into talent acquisition — sourcing, screening, scheduling, and increasingly, interviewing — the question Asimov was really asking is the same one talent leaders need to answer right now: How do we protect people from the technology people create?
After observing and experiencing various AI applications in talent acquisition, I’ve landed on three laws to guide AI use in talent decisions. While these are most directly relevant to talent acquisition, they apply broadly to any critical decision involving a human being! If you’re an executive who believes — really believes — that talent is your greatest asset, they’re worth adopting.
Law #1: Humans deserve to interact with humans. AI will not interview your candidates.
There’s a telling contradiction at the heart of many companies. Leadership will stand in front of an all-hands meeting and say with genuine conviction, “Our people are our greatest asset.” Then they’ll let a chatbot or AI tool be the first thing a prospective hire encounters.
Think about what that communicates. Every human touchpoint in your hiring process is a signal to the candidate about who you are and what you value. When that first touch is automated, you’re not just risking a bad candidate experience. You’re broadcasting, before the relationship even begins, the perception that you don’t value people.
The competition for top talent in most sectors is brutal. The executives and operators people want the most have numerous options. Forcing them to work through a bot as their first impression of your company is a fast way to lose them — and a slow way to build a team. While some might argue that the talent wants to move quickly and this is the only way to keep up, have you considered who are the people who you’re getting with this method and who are you losing? Who is self-selecting out? Would you interview with a bot in the same manner as you’re expecting of your new hires?
AI has a legitimate role in talent decisions. It can help you source more efficiently, schedule faster, and process more information. But the interview — the first real human encounter with your organization — must be led by a person. Full stop.
Law #2: Human decisions require human decision makers. AI will not decide for you.
This one is both philosophical and practical, and the practical case is getting stronger by the month. Philosophically, if people decisions are the most important decisions a company makes — and in my view, they are — then people need to make them. Delegating a hire or no-hire decision to an algorithm isn’t efficiency, it’s abdication.
But setting the philosophical argument aside for a moment, the practical risks alone should give any executive pause. We’ve all read the stories about AI systems that hallucinate, fabricate credentials, or surface confident conclusions built on nothing. We’ve seen the research on how bias gets baked into AI models — not always intentionally or with malice, but through the assumptions embedded implicitly by the person who wrote the model and the data that model was trained on.
When those biases touch hiring decisions, it doesn’t just risk making a bad call. It risks potentially making a call that can land the company in legal trouble. Notably, California and Colorado have both recently enacted legislation that turns up the heat on companies who rely on AI to drive decisions and clarifies the liability that they’re taking on.
Decisions that shape people’s livelihoods deserve better than that. AI can inform judgment, but it can’t replace it.
Law #3: AI will be used to both answer the question and negate the question.
Here’s what you need to understand about AI: it wants to please you. Ask it to build a case for a candidate, and it will build one. Ask it why the current process is working, and it will tell you. AI is extraordinarily good at confirming whatever it’s implicitly being asked to confirm. That’s not a bug. It’s how these systems are designed. In a hiring context, it’s dangerous because what you need isn’t confirmation — you need the truth.
The fix is simple. When I use AI in an assessment context, I’ve learned to ask it to argue both sides. Give me every reason this candidate is exceptional. Now give me every reason they’re not. Support both sides with specific evidence (interview notes, resume claims, testing results, etc). Once those analyses are complete, then a human evaluator can review, thinking critically about the strength and relative importance of the data, and come to a better decision.
Sometimes the first list wins. Sometimes the second list changes everything. Either way, you’ve used AI the right way — as an advisor that surfaces the full picture, not as a validator that tells you what you want to hear.
Asimov’s laws were written for fiction, but the instinct behind them was serious: in order to safely use powerful tools you need governing principles, established deliberately.
AI’s role in talent decisions isn’t going away. It offers clear benefits to help drive more structure and more robust data analysis into the process while also making it easier for the people running the hiring process and interviewing candidates. But the ones who get the greatest benefit from the technology will be those who use it judiciously, governed by the three laws above. The goal isn’t to see who can use AI the most or the quickest. The goal is to use it the best.
Post Views: 191
