What Legal Experts Need HR to Know About AI: Risks and Hazards

So let's talk through some things that HR may not be even thinking about.

What is one AI-related legal risk HR leaders consistently underestimate that could be putting their organizations at risk? And let's start here and we'll go that way.

Okay.

From my perspective, it's a lot more about, I'll call it shadow AI, this idea that, okay, well, we're doing something, but we're just not telling you.

so you know you have that employee that reads the policy but then decides yeah I'm gonna do it anyway or we have a proprietary AI but it doesn't give the results that I want so I'm gonna go over here to chat GPT or Gemini and do it a different way and if I if I do it in some way that doesn't make sense but it's not a big deal and that's much more of a nightmare I think then people realize

Yeah, I'm going to start off really basic.

And it's about use case.

I think one of the biggest risks that folks are underestimating is that AI doesn't fall into a category where one size fits all.

And so we need to understand what the tool was intended, what it was developed to do.

We need to put controls around how the tools are used in our workforce and then finally i would say there's a lot of hedging in terms of you know okay we purchased this tool for xyz hiring right but we're not really using it to make any final decisions and certainly the the vendors or the developers of the tools will often state This isn't meant to be used to make selection decisions.

Those decisions are the employers to make.

And I feel like employers are starting to pick up on that theme and say, maybe there's a possible like defense strategy here where we could simply say we're not really using it to make decisions.

So it's not a tool that we should be concerned about.

And in reality, I think that is going to find a lot of employers in hot water.

because, of course, we can use statistics and we can use data and we can use outcomes to understand how tools are actually being used.

I'll try to be brief, but two points to make.

One, to echo what Michelle said and to amplify it.

When you are told that the final decision is not being made by AI, there is no relief because every decision to screen someone out or to disadvantage someone and advantage someone else by giving someone a more direct path to an interview.

Each one of those decisions are the decisions that the laws are looking at, because they're looking at it from the point of view of the person who was denied, not the person who was selected.

That's why there is no such thing as a final decision.

Every decision matters.

To answer your question, Nicole, on additional risks that HR leaders should be cognizant of, I think HR leaders in the world in general should be aware that AI is not really as intelligent as we tend to assume it is because perhaps Asimov wrote some really good novels about that.

The fact of the matter is that there's a guy named Andrej Karpathy who talks beautifully about AI and I'm about to take one of his points.

The way these AI systems were developed were to mimic the human brain And in short, they haven't gotten anywhere near the depth and complexity of the human brain.

You hear terms like neural networks.