“How do I know the AI tools I’m using are safe, accurate and approved?” ADP’s Kasara Weinrich and Helena Almeida discuss top questions to ask when using AI at work. For more, visit ADP.com/SPARK

Video Transcript:

Kasara Weinrich

how do I know the tools I'm using are safe, accurate and approved.

Helena Almeida
Yeah. So that's a that's a great question. It's a good question to be asking. Right. And you should always be asking that question. Most organizations have or will soon have kind of an approved list of tools like this is the approved list. We've checked. We've checked the security. We've we understand how it works. And we sanctioned this for our workers to use in connection with their with their work.

I know that on our end, we definitely get a lot of organizations that are asking us about our AI because they want to make sure that they can put it on that list for for their workflows. Ask questions. I mean, I think, again, sort of like like the person who was worried about it making a mistake. Don't not use anything because you're worried you're going to use the wrong tool, right?

Ask questions. It's important for organizations to be clear to other workers, because if you want people to get benefit out of the tools that you're investing in, you really need to kind of clarify the rules for them. And and those that clarity is going to help both ways. Right. So ask questions if something's not clear so the organization can clarify it for you. And that's the best advice I would give to that worker.

Kasara Weinrich
Yeah. And it does make sense to from you know depending on the industry that you're in, depending on the size business that you work for, you're likely to receive varying levels of guidance around what is safe or what is approved. I think to one of the things I haven't necessarily seen in a lot of messaging from many employers is any kind of guarantee or validation around accuracy, like this worker specifically asked, how do I know if it's safe and accurate?

And I think it does go back then to the conversation around education and around the understanding of this is the approved tool. This is why this is what it's maybe connected to internally within our organization for our own intellectual property. I think this goes back also on the employer to make sure they're communicating appropriately.

Helena Almeida
You know, there are certain things that I think AI can be more accurate than than, for example, I'm not great at math. I know that's basically a stereotype for lawyers, but it's true. I'm not. And you know, AI or calculator is going to be better at math than I am, right?

I have judgment where I can say that number is way outside the ballpark of what I expected it to be. So sometimes I can rely more on a tool, a calculator, AI, you know, spell check to probably be better and have a better output than I would have come up with on my own if I. My spelling is is okay.

On the other hand, I am better at certain things than than AI will be. So it's always going to be sort of. There's no way to like absolutely measure accuracy. What I can say, though, is that if you know enough about how the tool works, you might learn that the tool is pointing to accepted sources. Right? So for regulatory information that we might provide our compliance rules, right? We we will make sure that we're curating the information that that comes from. So it's not kind of a random person kind of feeding, feeding their own feelings about payroll into the internet. We're actually kind of citing to really trusted sources to make sure that the information is accurate as possible, but there's always going to be kind of differences in the level of accuracy depending on what the task is and reliance on, on people to make sure that they're exercising their own judgment.

One of the things I'll say is that when we're trying to be as careful as possible, like there is, there's not a lot of wiggle room when you're working on someone's payroll. There's not a lot of I feel like it was close enough. So there are different tools that vendors can use to make sure that you're being as accurate as you possibly can in some circumstances. And there are other circumstances where you might want more flexibility in the answer and a little bit more a little bit more creativity from the AI.

So, you know, the great thing about what we do is we're trying to figure out what the right tool is for the right situation and, and making sure that we're we're building in our experience and kind of expertise into the product so people can have a little bit more confidence in what the output is.

Kasara Weinrich

It's amazing. And it's something that then employers don't necessarily have to worry about, and they can have a lot of confidence in saying, hey, worker, this is obviously on the approved list because ADP has already done the work.