Transcript
Nycole Rosen (27:17.993)
Yeah.
Matthew Mottola (27:30.193)
Where do you see AI going from like a how can it really ruin companies? and I asked that as the high level because I think governance is kind of a subset of like, well, what if you just let all your employees run nuts on Claude? you'll see what happens. But but what can go wrong? Give us give us like the horror stories. Yes. Give me the demon glimpse, which which I'm not, I'm not, what's his
Nycole Rosen (27:49.039)
man, you want the doom and gloom. That's what sells, right?
Matthew Mottola (27:56.308)
What's his name? Dario will come up and say, the world's gonna break, everything's whatever. But then he'll but then he'll wink and be like, by the way, please regulate all my competitors. And it's like I've I've been in tech long enough, I know what game you're playing. but so yes, give me give me the full doom and gloom and then say you have to regulate everyone besides you. Just kidding.
Nycole Rosen (27:58.651)
yeah. It's gonna kill us all.
Nycole Rosen (28:07.47)
Yeah.
Nycole Rosen (28:13.484)
Wow. so I'm sure that we've all seen the headlines that have come out recently of the AI agents that got out of their sandbox and started doing things they shouldn't have done. I'm not gonna name names. or or people that believe that AI could kill us all. And I I feel like those are provocative and that gets attention, but I don't know that I lose sleep about any of that. because
It's going to take a few of those examples, unfortunately, and then we will get some regulation. Right now we're not as a as a society heavily regulated with AI usage. And I think that's for two reasons. One, because it's so new and we're just unsure how to regulate it. But also there's always hesitancy to regulate, unfortunately, until there's a reason to do it. Pretty much every law out there requires.
became a law because something happened. we have speed limits because people were driving too fast and running into each other. We have to wear seat belts because we started to see an increase of people passing away from car accidents, for example. So unfortunately, I think we are going to have to see some significant problems come out of AI before we get a lot of regulation.
So to answer your question, what could happen? And I'm going to stick with the staffing industry because that's what I know, but I think we could all let our nightmares go wild and think about what AI could do. and and there's plenty of that out there. So if that's what you're looking for, I'm sure you can find it. the things that I think are more realistic that could go wrong in staffing is you let AI take over making decisions that require a human.
And your AI will start to lean towards certain decisions and start becoming more and more biased to a point where it's doing extreme discrimination and leaving out a huge subset of candidates and only hiring people within a certain category. And that will be learned over time. It's highly unlikely that a person is going to go tell at that and say, only hire this type of person and leave out everybody else.
Nycole Rosen (30:25.698)
But it'll start looking at the data and it'll start r noticing that, okay, we're only hiring this type of person. So if someone else applies that's outside of that category, we're going to automatically exclude them. Of course, we know that there are anti-discrimination laws out there. So that seems to be the easiest one that we notice. And we've actually seen that happen inside and outside of the staffing industry. And organizations see that, hopefully, and get ahead of it quickly.
that could happen. We could start seeing it make decisions that we wouldn't want it to make at all if we give it too much freedom. And I understand you you've connected yours to your bank and your everything out there. And I think we all kind of do that a little bit, but we in staffing have a responsibility. Just like in healthcare, healthcare providers have the the HIPAA guidelines that they have to follow. Staffing, we have a lot of people's personal information.
that we are responsible to keep private and protected. And if we start allowing our AI to have access to all of that, we are opening up not ourselves, but all of our database of candidates and maybe even customers with their bank information. And then we are opening ourselves up for potential litigation and and wrongdoing. And we certainly don't want to do that.
Matthew Mottola (31:49.458)
Yeah. No, I I
I I totally agree. I think what's interesting about AI, and we I mean we published our book in twenty twenty before LLMs had hit. And so, you know, I've never forget my my co-author, who was like CTO, CIO, like built some of the early deep learning models. and the way he described it is he was like, it's just an amplification effect. He like like still, he was like, all the bad things that can happen, there's AI doesn't create any really net new things, it just simply amplifies and accelerates.
Rights, all the things that already can go wrong, whether it's jailbreaking, whether it's discrimination, all those things. and we s we s we've seen it, right? Like in early models, like it was it Google, like plenty of them have got in trouble for having these models that was like, okay, that that reflects all the biases that that culture has, right? So it makes total sense. Now, I wonder, like, last sort of last question for you is.
Nycole Rosen (32:42.562)
Yes.
Matthew Mottola (32:50.213)
When you when you're talking to a company, what is the one thing that you want them to ask themselves? Because I also think this is where we see this big disconnect where you have people like me who are like early Silicon Valley tech person, does believe AI is gonna take over the world, not in a bad way, just believe that it you know, it'll be in you know, in everything.
