top of page
Indeed Wave.PNG
DS_logo_Primary.png

UNLEASH Wrap & A.I. Hiring Regulations


Live from Paris at UNLEASH World, the boys are on stage with Keith Sonderling, EEOC Commissioner & Oana Iordachescu Associate Director, Talent Acquisition - Technology Europe/Asia at Wayfair Deutschland. And they’re talkin’ A.I. issues in hiring, focused on Europe. We’re always keeping you away from land mines, no matter the continent. What’s more? How about a special show summary with fellow podcaster Matt “That British Guy” Alder? You’re welcome.


INTRO (1s):

Hide your kids! Lock the doors! You're listening to HR’s most dangerous podcast. Chad Sowash and Joel Cheeseman are here to punch the recruiting industry, right where it hurts! Complete with breaking news, brash opinion and loads of snark, buckle up boys and girls, it's time for the Chad and Cheese podcast.


Matt (26s):

Press record now.


Chad (33s):

You are mispronouncing it. It's called shit.


Joel (40s):

Oh yeah. What's up everybody? We are at Unleash World. You know the game. This is the Chad and Cheese podcast.


Chad (48s):

Gay Paris.


Joel (48s):

I'm your cohost, Joel Cheeseman. Joined as always, "the Laphroaig to my Aberfeldy" Chad Sowash. And we are, if you can stop laughing. Welcome Matt "that British guy" Alder popping in for a visit. Taking a break from podcasting. Matt, welcome to the Chad and Cheese podcast.


Matt (1m 9s):

Thank you very much. I can hear the whole of Scotland now up in arms about the way that...


Joel (1m 13s):

How many kilts just went up in flames?


Matt (1m 16s):

Mispronouncing some of their greatest love brand names


Joel (1m 21s):

Fellows.


Chad (1m 23s):

Aberfeldy


Joel (1m 24s):

Prince Aberfeldy.


Chad (1m 25s):

Really don't enunciate things in Scottish. Right. Everything just kinda like runs through. But it's got this wonderful.


Joel (1m 31s):

You're offending him. Look at how red's getting, stop offending it. It's a


Chad (1m 33s):

It's a very rich kind of like robust guttural.


Joel (1m 37s):

Yeah.


Chad (1m 37s):

Love is what I feel when I hear somebody speak.


Joel (1m 39s):

I've never been. But we're gonna change that.


Chad (1m 41s):

But we know Scotts. So


Joel (1m 43s):

We know Scots. We like all the Scots. We know it.


Chad (1m 46s):

You're really Welsh though? Right?


Matt (1m 47s):

English? I'm English.


Chad (1m 48s):

Oh, I thought you were Cardiff. I had Cardiff by the sea. No?


Matt (1m 50s):

No, no, no, no, no. I'm English. Okay. I used to in Cornwall, but that's part of England.


Chad (1m 55s):

Sorry about that.


Matt (1m 56s):

So, I now I live in Scotland. Scotland, okay. Yeah. So I'm not a Scottish, but my wife and family are all Scottish. Oh yeah.


Chad (2m 2s):

But you live in Scotland. So


Matt (2m 4s):

I'm becoming assimilated slightly. I'm not wearing a kilt, yea.


Joel (2m 7s):

They accepted you yet?


Matt (2m 8s):

Just very open, accepting inclusive country. So you'd even be welcome.


Chad (2m 12s):

I don't know about that.


Joel (2m 14s):

Wow. That's saying something, cuz I offend everybody.


Chad (2m 18s):

Yes. We're here today kids, in liEU, let's say of the Friday show. We're gonna go over a Unleash recap, talk a little bit about all the fun parties. People are back. We're in Paris. But at the end of the day, listeners want to hear how this shit went. Yes. The FOMO is real. Kids.


Joel (2m 36s):

Well, fomo all over like we're in EUrope. Fomo. We're Paris.


Chad (2m 40s):

Is that a song?


Joel (2m 41s):

We're at Unleashed? It could be a song. Let's write a song right here, right now.


Chad (2m 46s):

Fomo all over. I'm gonna FOMO all over. No, sorry, go ahead.


Joel (2m 51s):

I'm going to get the Vonqin' FOMO on you. It's great to be back in person. That's all I have to say.


Chad (2m 59s):

It is. Yeah.


Joel (2m 60s):

We're human beings. We need to socialize. And for two and a half years we didn't get to do that. Especially as podcasters, we talk to people, we speak into the ether to no one and it's great to come and see people.


Chad (3m 13s):

Yes.


Joel (3m 13s):

I don't want to break my own arm patting myself on the back, but to have fans come by the booth, I'm sure you've experienced the same thing. Like, you know, love you guys keep, keep doing what you do.


Chad (3m 25s):

Matt's big in Europe.


Joel (3m 26s):

You're full of shit on my startup that you talk shit about. It's great to get all that feedback.


Matt (3m 31s):

I Think that's well it's great to have. We've got this podcasting area with these kind of branded backdrops.


Chad (3m 37s):

Pretty sweet. It's pretty good.


Matt (3m 38s):

Yours is kind of bright yellow.


Chad (3m 40s):

Ours is like screaming.


Matt (3m 41s):

The people coming by my booth have mainly been coming to plug their phones in to recharge cuz I've got a spare socket.


Chad (3m 47s):

We covered ours up so people couldn't see it.


Matt (3m 50s):

Yeah, absolutely. But it's, no, it's a really, it's a really great event. I mean there's lots of interesting companies exhibiting. I was hosting the recruitment stage yesterday. There's some great content.


Chad (4m 1s):

Yes.


Joel (4m 1s):

Tell the listeners what that is. What kind of content were you overseeing? What were they?


Matt (4m 7s):

So it was a string of very senior TA people.


Joel (4m 10s):

Okay.


Matt (4m 11s):

Talking about their various challenges. We had a great panel on first up talking about all the things that are affecting talent acquisition at the moment and how large companies are dealing with them. We heard from Deloitte, we heard from Deutsche Bank all kinds of other, you know, large, large EUropean or large multinational companies looking at all kinds of different areas of Talent Acquisition.


Joel (4m 34s):

They all say we're in a hiring freeze for the next 12 months.


Matt (4m 37s):

No! Not at all. There was still.


Joel (4m 39s):

Actually activity.


Matt (4m 39s):

Yeah.


Joel (4m 39s):

Cuz there was a lot of Americans think it's doom and gloom over here in EUrope and not so much.


Matt (4m 43s):

I mean it is but that doesn't translate.


Chad (4m 45s):

Doesn't stop us.


Joel (4m 46s):

But we're still hiring.


Matt (4m 48s):

It's still really the sense was that, you know, there's lots of economic pressure recession in quite a few countries, recession coming into ones that aren't there yet.


Joel (4m 57s):

Yep.


Matt (4m 58s):

However, still very difficult to attract and retain talent. So, you know, companies may not be hiring in the numbers that they were hiring before, but they've still got all the same problems they had a year ago. And so interesting to see people linking TA also and employer branding with retention. So retaining people still big issue. How do you build


Joel (5m 21s):

Upskilling.


Matt (5m 21s):

cultures? Had you upskill people all kind of thrown in the same presentation, which we didn't see before. That would be three different presentations, like two or three years ago. So, that was really interesting yesterday to see that.


Chad (5m 34s):

Pretty amazing. I mean we see the big names. They're all over the expo floor here.


Joel (5m 39s):

Not all the big names.


Chad (5m 41s):

All of the brands that are now, especially cuz we're in EUrope that are the Atlases, the Remote, the EOR companies.


Joel (5m 50s):

Yeah.


Chad (5m 50s):

Who are popping up and like exploding. Well probably because they got hundreds of millions of dollars, but exploding because EUrope, there's a bunch of countries and it's not easy to hire in EUrope for all of the different countries. But when you're looking for talent and you have the problems that you were just talking about, Matt, you need that those types of services. So we did see them in the US at HR Tech, but they, I mean they're big and strong here. Talk about that a little bit.


Matt (6m 20s):

Yeah, absolutely. It, you know, again, going back to the speakers that I saw yesterday and the people who've come on my podcast that I've sort of interviewed over the last couple of days, you know, they're all recruiting in multiple countries. All kinds of issues with, you know, different regulations. Lots of different things going on.


Chad (6m 37s):

Yeah.


Matt (6m 38s):

And yeah, just an explosion of tools, you know, to help that and also to really help people tap more into, you know, global workforces and, you know, borderless working and all that and all that sort of stuff. So yeah. Really interesting.


Chad (6m 53s):

We have a drive by listener


Joel (6m 56s):

A drive by.


Chad (6m 56s):

Who are you sir?


Lieven (6m 58s):

I tried to stay anonymous.


Joel (7m 2s):

We don't do anonymity.


Matt (7m 4s):

But we can see you.


Chad (7m 6s):

So Lieven. So yeah, Lieven is here, but for some reason he's been trying not to get on the mic, so I had to get him on the mic. So about this show, other than not going to the champagne tasting. Well what's been great?


Lieven (7m 20s):

The exhibition basically it's marvelous. So many companies I never heard about and I'm trying to hear about now.


Joel (7m 27s):

Lieven's making a shopping list for House of HR acquisitions in 2023.0.


Lieven (7m 31s):

Buy this and this and this and we're gonna buy all something like that. And I was trying to build the best I recruitment congress in the world and now I think have some work to do.


Chad (7m 41s):

Oh. which is going to be where next year?


Lieven (7m 46s):

Amsterdam


Joel (7m 47s):

Hold on. So did I heaer him the eCongress is gonna be this size at some point. Like you have bigger ambitions than what it is?


Lieven (7m 58s):

Smaller yet better.


Joel (7m 60s):

Oh. Kinda like Belgium. Smaller but better. Matt, liked that one. Lieven's taken us to Moulin Rouge tonight. We need to be nice to Lieven.


Chad (8m 8s):

I need to work through that. Yes.


Joel (8m 11s):

Boobies and Champagne tonight apparently is what was on tap. It's France.


Chad (8m 15s):

It's France. So that should be standard. So yeah. So we started off with the Vonq rooftop party. matt, you were there. You were there.


Matt (8m 22s):

I was there.


Chad (8m 22s):

You were there.


Joel (8m 22s):

You were there. You got in a day before. Yep, I got in the day of the party. Did you come in on Tuesday?


Matt (8m 29s):

I literally came in, I landed an hour before the party started. I came straight to the party.


Chad (8m 33s):

So talk about last minute. I mean, Jesus Cheesman comes in like a few hours before. Then Alder makes.


Joel (8m 41s):

I ride in, I make my plans well in advance. The show starts on Wednesday. I come in on Tuesday. That gives me like 12 hours.


Chad (8m 51s):

Gotta know that something is happening prior.


Joel (8m 52s):

Get adjusted prior and then a month before a party gets scheduled or hey, somebody's doing it. So then it's like,


Chad (8m 59s):

It's Paris!


Joel (8m 60s):

I got two hours.


Chad (9m 1s):

It's Paris.


Joel (9m 1s):

Instead of 12 hours.


Chad (9m 3s):

You got a little nap. Yeah,


Joel (9m 4s):

Unfortunately I have a five year old, it's harder for me to,


Matt (9m 7s):

But you know, I'm on an hour time different. So I didn't really need it. Didn't really need a nap.


Chad (9m 12s):

Okay. That's good.


Joel (9m 15s):

That's good. And Matt always parties all the time. So he's ready to go.


Chad (9m 19s):

Yeah.


Joel (9m 19s):

At a moment's notice.


Chad (9m 21s):

Always. So that was a blast. Rooftop.


Joel (9m 25s):

Dude. Paris rooftop. Eiffel Tower. Sun setting. Perfect weather.


Chad (9m 29s):

Free drinks.


Joel (9m 31s):

Perfect weather. Yeah. Free drinks. French whiskey and fried chicken.


Chad (9m 35s):

French whiskey. Surprising.


Joel (9m 36s):

Surprisingly good.


Chad (9m 37s):

I could not believe the fried chicken.


Joel (9m 41s):

Some fried chicken at this party.


Chad (9m 42s):

You know what it tasted like? It was from some upscale shake and bake. That's what that shit was.


Joel (9m 47s):

Yeah, you're right. Chicken breast. And there was some steroids in these chickens cause they were huge. Chopped nicely. Breaded exquisitely.


Chad (9m 55s):

It was shake -n - bake.


Joel (9m 56s):

A nice little dip.


Chad (9m 58s):

Yep.


Joel (9m 58s):

It was fantastic. And the French whiskey, I'm kind of convinced that it's Irish whiskey aged in some sort of French barrel.


Chad (10m 5s):

New label.


Joel (10m 6s):

Yeah, it's made in Bush Mills distillery. And then they aged over here.


Chad (10m 9s):

See I got Matt one of those and he was like, I'm not drinking that shit. I'm like, Matt, really? I'm like, Matt, Matt, try it.


Matt (10m 16s):

I really wanted a beer. That was why I really wanted a beer. They had a very nice IPA so it was good. But actually the whiskey was very nice. It'd be nice to know what it actually actually was.


Joel (10m 30s):

It was Irish whiskey?


Matt (10m 31s):

Irish generic French whiskey. Yeah, Maybe, Maybe.


Joel (10m 34s):

Maybe.


Matt (10m 34s):

Yeah.


Joel (10m 34s):

Definitely not scotch and definitely not bourbon.


Chad (10m 36s):

Say it was. Yeah, I know it was, it was a little, it was much smoother.


Joel (10m 39s):

It was light. Triple distilled for sure.


Chad (10m 43s):

Yeah. More of a sweeter.


Joel (10m 44s):

Yeah. So thanks to Vonq for having the party. Yeah. Supporting, supporting that. After that we had a nice Italian dinner that you took me to in Paris. Cuz that's what you do in Paris. You have Italian food.


Chad (10m 56s):

When you're going somewhere with Julie Sowash, that woman wants pasta. I mean, and she hadn't had pasta for a couple of days, so we had to make sure that she got it.


Joel (11m 5s):

She loves the noodle.


Chad (11m 6s):

Yes. The rigatone noodle. So yeah, no, that was a good time. Plenty of wine. Plenty of good time. The first day of the show got in here. I was really expecting to see more of kinda like a downsize because you know, people coming back. Is it gonna come slower? It didn't.


Joel (11m 22s):

This one might be a tad smaller, a tad more intimate. No?


Matt (11m 26s):

I think it's a similar size to the one there were three years ago. Yeah. I mean it's difficult to remember anything that happened before the pandemic. But it seems familiar at this kind of size.


Joel (11m 35s):

The Eightfold booth is smaller. Maybe that's why they didn't go with the 380 by 380 booth at this one.


Chad (11m 42s):

Maybe, they actually limit.


Joel (11m 43s):

Oh that would be very European.


Chad (11m 45s):

Yes. The amount of booths that you can have in one fucking space.


Joel (11m 49s):

Don't be too American with your booth size.


Chad (11m 51s):

No. Then we did dinner with Max from Talk Push last night.


Joel (11m 56s):

Yeah. Well we did a session with our friend Keith Sonderling, which was great. And that was well received. A good crowd.


Chad (12m 3s):

Which you're gonna hear right after this short update.


Joel (12m 6s):

Which will play for you after this. But yeah, funny story you, you took off to get ready for dinner, which we had with Max. I had the intention of going to the Eightfold party talking about Eightfold. And there are apparently two hotels of the same name that both have a rooftop bar. I went to the wrong hotel. I did go to the rooftop bar and I go up and I say Eightfold party and they spoke French so they didn't know what I was saying. Luckily I saw Sarah White and Lagunas.


Chad (12m 37s):

Kyle Lagunas?


Joel (12m 38s):

Yeah. In the corner there. And I said, yes, this is where I belong. They let me in. I had some Laphroaig.


Matt (12m 45s):

You got it right! You got it right.


Joel (12m 47s):

And it was great. I met some new friends and then left and met up with you at a real French Cafe, Bistro restaurant, whatever they call it.


Chad (12m 55s):

It was very small. As a matter of fact, Max Arm Brewster was actually spent a good amount of time.


Joel (13m 4s):

What did he have? Sheep brain?


Chad (13m 5s):

Some brain something.


Joel (13m 5s):

Gorilla Brain. That was a weird French.


Chad (13m 7s):

French will eat everything. Apparently. Apparently. They'll eat everything.


Joel (13m 10s):

Drink and eat everything.


Chad (13m 12s):

Yes. But yeah, to Talk Push thanks to Talk Push for dinner last night. Woke up this morning. I felt like shit not because of the alcohol.


Joel (13m 18s):

Yeah.


Chad (13m 19s):

Little bit of a cold.


Joel (13m 20s):

Yeah. You're dragging a little bit.


Chad (13m 23s):

Little bit. Little bit. But other than that,


Joel (13m 24s):

Chad dragging is like everyone else's 100%. So it's not too bad.


Chad (13m 30s):

And a shit ton of interviews we had Zev from Syndio. Oana was on stage with us. Yeah.


Joel (13m 39s):

Fuel 50, Textkernel. VONQ. Yeah. It was a parade of influencers and companies making miracles happen is what we had.


Chad (13m 44s):

Yeah. And we got some Wayfair too, so that's good.


Joel (13m 47s):

They got just what we need, Chad.


Chad (13m 48s):

Yes.


Joel (13m 48s):

That Wayfair.


Chad (13m 49s):

And then we got Bas on and we had to cut him to 10 minutes.


Joel (13m 55s):

I don't know. We need to get Bill Borman and Boss to go bowling.


Chad (13m 59s):

Jesus Christ.


Joel (13m 60s):

The eight hour bowling match is what that would be good.


Chad (14m 6s):

Yes.


Joel (14m 6s):

God, those guys could talk. A Brit and a Dutchman go into a bar and stay there for 24 hours and talk apparently.


Chad (14m 13s):

Any parting shots, Matt?


Joel (14m 14s):

Words of wisdom because they generally don't get that from our show.


Matt (14m 16s):

Yeah. Well I just, you know,I did go to the Eightfold party last night, so.


Joel (14m 22s):

Okay.


Matt (14m 23s):

I went to the right place. I read the instructions properly, but also I just wanna make a big shout out for Social Talent who had an amazing dinner last night. Possibly the smallest room I've ever been in in my life with the most amount of people.


Joel (14m 37s):

Did you say sexual talent?


Matt (14m 38s):

Social talent.


Joel (14m 38s):

Social talent.


Matt (14m 39s):

Social talent. Sorry, I think you're hearing is going.


Chad (14m 42s):

He's always hearing what he wants to hear.


Matt (14m 44s):

Absolutely. No, definitely Social Talent. So shout out to them cuz you know, that was some great conversations there.


Joel (14m 49s):

Tell the story about their booth or lack thereof. I think that's a funny story.


Matt (14m 53s):

I don't know that story.


Joel (14m 54s):

The booth next to us is having a barista and they have really nice coffee.


Matt (14m 57s):

Yes.


Joel (14m 57s):

So, Dave is over there. I'm like, Hey guys, where's your booth? Da da da. So apparently someone I don't know, forgot to like, have the booth delivered if there is even a booth. Who knows? So they have no booth over there. They've totally just like gone with it. Fortunately everyone there wears these green like super green shirts with the ninja thing. So Yeah, that's my funny story. They have no booth, but they have a booth and they're making it work like truly Irish.


Matt (15m 25s):

Leaning into not having a booth.


Joel (15m 27s):

Yeah. Like a true Irish, like let's just roll with it. No booth.


Matt (15m 32s):

Yeah.


Joel (15m 33s):

Which is a great conversation starter.


Chad (15m 35s):

Which I love is what Johnny said. He said, it sucks, but nobody died. Let's make this work. And that my friend is how this shit should work in the first place.


Matt (15m 43s):

I think that should be the motto of the next 12 months.


Chad (15m 49s):

Amen.


Joel (15m 49s):

Chad's words to live by. Take a step back, take a deep breath and ask yourself, did somebody die? And if they did, call 9 1 1. But generally no one died.


Chad (15m 59s):

What is it here? Is it 211? 112 or something?


Matt (15m 60s):

Anyway it's different.


Chad (16m 5s):

You can gone on that one.


Joel (16m 8s):

Nine one. Somebody call 211. Cuz this podcast is dead. We out.


Chad (16m 13s):

We out.


7 (16m 14s):

Good afternoon, guys. Welcome to the Influencer stage. You'll see that this is down as a panel, but the two guys that lead this, obviously the most hrs most dangerous podcast, never have a panel. Plus it definitely is a fireside chat. And they will be unpacking. Can you trust hiring with AI? Over to you guys.


Chad (16m 39s):

Hello Paris?


Joel (16m 40s):

Oh Yeah.


Chad (16m 40s):

Hello. Where's the free, where's the free champagne? That's the question. Everybody should be getting free champagne at this point. I am Chad of the Chad and Cheese podcast.


Joel (16m 50s):

I am Cheese, Joel Cheeseman of the Chad and Cheese podcast. If you don't know about us, you can check us out on any podcasting platform that you currently enjoy. And we are just giddy, giddy as hell to introduce these two guests with us. First and foremost, Oana Iordachescu. Did I butcher that too much?


Chad (17m 10s):

Oh yeah, that was too much. That was horrible.


Joel (17m 16s):

She's associate Director of Acquisition Technology in EUropean Asia at Wayfair.


Chad (17m 18s):

Wayfair.


Joel (17m 18s):

Wayfair. If you haven't gotten your furniture at Wayfair. What are you waiting for?


Chad (17m 23s):

You gotta sing it when you say it.


Joel (17m 24s):

Wayfair, you've got just what I want or need. Want or need. I don't know. And you, go ahead and introduce Keith.


Chad (17m 33s):

Okay. So Keith Sonderling. Yes. A commissioner at the EEOC. We actually had bring the US over to have a conversation around EU.


Keith (17m 40s):

Taking the show on the road.


Chad (17m 42s):

Yes.


Joel (17m 42s):

America.


Chad (17m 44s):

Keith, give us a quick Twitter bio of you. Long walks on the beach.


Joel (17m 50s):

That's right.


Chad (17m 51s):

Frequent


Keith (17m 51s):

Thats's right, frequent Chad and Cheese.


Joel (17m 54s):

He loves a good perp walk. That's right.


Keith (17m 56s):

Hello. My name's Keith Sonderling, I'm commissioner at the United States Equal Employment Opportunity Commission. And what that means is we are the premier civil rights agency in the United States, responsible for enforcing all workplace discrimination laws. So everything from the "Me Too" movement, to pay equity to disability, you name it. Everything relevant to HR professionals. That is my agency.


Joel (18m 17s):

And a Twitter bio from Oana. Is that enough characters? That's good for now. That's good for you.


Oana (18m 23s):

You do all the cool stuff.


Chad (18m 26s):

He's in government. You just said he does all the cool stuff. Really?


Oana (18m 31s):

Yeah, I mean, Me too. Discrimination. Lawsuits. I love that! Oana Iordachescu for those who want you're the try harder next time.


Chad (18m 43s):

Oana Iordachescu.


Oana (18m 44s):

I live currently in Berlin. I'm originally from Romania and worked in technology recruitment for the past 10 years with booking.com. Facebook always building technology teams in this organizations, of course as part of TA we use a lot of these tools that they are trying to regulate. And that's a good thing. So we're gonna talk about that.


Chad (18m 60s):

Oh, regulation's a good thing. We must be in the EU.


Joel (19m 5s):

She kept it professional, didn't she?


Chad (19m 8s):

We must be.


Keith (19m 8s):

I don't hear that often in America. Here get away from us.


Chad (19m 12s):

Yes.


Joel (19m 13s):

Oana also loves disco, dancing and dressing like a rock star. Let's get to the Q and A, shall we?


Chad (19m 19s):

Let's do this.


Joel (19m 20s):

So the AI technologies that many attendees are seeing here today are filled with a minefield of potential risk.


Chad (19m 28s):

Are you saying risk?


Joel (19m 30s):

Risk? I'm saying risk. Risk is out here everywhere. Keith, what are some prevalent risks that people should be looking for, from an EEOC perspective as they're talking to these vendors around the expo hall?


Chad (19m 38s):

Even here in the EU?


Joel (19m 38s):

I mean, yeah, and I wanna take a step back and it's really important to, you know, put the context of where we are and how we got here. So artificial intelligence, as we all know, it's a buzzword. It's throughout the industry. So the other sides of your businesses outside of HR have probably been using artificial intelligence to make your business more effective, efficient, make more money. And there's been so many products to out there under the AI side that your C-suite and your board of directors have loved because they've implemented that. It's saved costs. It's just made companies more profitable. So now we're seeing AI coming into a different area, which is our space and human resources with the promise to eliminate bias.


Joel (20m 23s):

And I think that's a really thing.


Chad (20m 25s):

Promise.


Keith (20m 25s):

With a promise. Promise to eliminate bias because you have these very smart computer engineers who have figured out to make the other size of your business better. And now let's make HR better. And how do we do that? What do we have to eliminate? Now I'll ask the questions, Chad and Cheese. What do we have to eliminate to make?


Chad (20m 44s):

That's not how our show goes.


Keith (20m 45s):

An AI product, super effective. Eliminate bias, eliminate discrimination? The H or the R? Which one do we wanna eliminate?


Chad (20m 49s):

Probably the humans. Yeah, we don't need the humans right?


Keith (20m 53s):

Why do we wanna eliminate the humans?


Joel (20m 54s):

They're a pain in the ass.


Keith (20m 56s):

They're a pain in the ass, but they're also, you know, full of potential bias. So in all seriousness, where a lot of these products are being sold is to eliminate the human from the decision making process. And if you eliminate the human, you also eliminate some of the longstanding biases that have plagued the workforce. The reason my agency exists in the United States and similar agencies here in the EU, is because employers discriminate. And whether they intend to do it or not, they're going to be liable for those decisions. Both from a government enforcement perspective and a publicity's perspective of discriminating against women, discriminating against agents, you name the protected characteristics.


Keith (21m 36s):

It's not good news for any of that.


Chad (21m 39s):

So getting back to your question, artificial intelligence out here, and a lot of vendors are now creating a product to eliminate that bias. To just use nEUtral characteristics, whether it's the characteristics related to applying for the job, succeeding in the job, to make the products have less bias. Yeah. Does that sound simple? It does, but I don't, I don't think Oana buys that. Do you buy that?


Keith (22m 1s):

I'm not selling anything. I'm the government.


Chad (22m 3s):

Are you buying that, that it's eliminating risk?


Oana (22m 4s):

No, absolutely not. Okay. It's not eliminating risk. I think actually it's probably creating some interesting side effects. And a lot of these tools we're talking about, they're not actually being sold to us to eliminate bias. They are sold to us to make the process more effective, faster. Right? So I think the different products that we can use, promise that while whatever you're trying to solve, speed, quality, volume, whatever, it's not generating more bias than you would with humans, Right. Because you're liable if you do discriminate without the ai?


Chad (22m 34s):

With the humans.


Joel (22m 35s):

Yeah. Which is a great point because I think a lot of people believe, well if it's their product, am I on the hook for it? Because I didn't create it. It's not an Amazon situation where we, this is a homemade technology. So I think a lot of these people are here concerned about what questions should I be asking these vendors to make sure that I'm not in trouble when shit hits the fan?


Keith (22m 56s):

And in the United States, the employer is liable for the employment decision. So whether it's made by a computer or whether it's made by a human, you can't say, Well yeah, we discriminated, but the computer did it so we're off the hook and we can't get in trouble. That's just not how it works. And I don't think anywhere in the world they're going to buy that. So the challenge is now with all these very innovative products coming online that truly, and we'll talk about if they're designed properly, whether you know it's the right data, the very diverse data set to help you recruit, or if it's an algorithm that that doesn't let you filter based upon somebody's sex, their national origin, it can actually help companies with their diversity can actually eliminate some of the bias.


Keith (23m 38s):

But if it's not, it can scale discrimination larger than we've ever seen before. So there's always two sides of the equation here. But you know, that oversight is what's gonna be critical.


Chad (23m 50s):

So to Oana's point is that humans are biased. There's no question there, they're biased. But guess what humans can't do well? Scale. So when you start using ai, if the AI is biased because of human developments, guess what? You're gonna scale bias, which means your opportunity to be more biased goes up dramatically and it goes up Fast.


Keith (24m 12s):

Yeah. And there's a stat that the average talent acquisition person takes around seven seconds to look at a resume, right?


Chad (24m 17s):

Yes.


Keith (24m 17s):

That's six. Okay, we lost a second. And that's not because of, you know, they have short attention spans, right? It's because of sheer volumes of resumes you're getting.


Chad (24m 25s):

Yeah.


Keith (24m 25s):

So think about if you have an HR manager or talent acquisition manager who doesn't wanna hire older women, right? And so they have to look through, okay, when do they graduate college, what's their name? Okay, here's one in the trash. No here's the trash. With an algorithm in 0.7 seconds, probably even faster, you can scale discrimination from what that one person takes the look through resume in a millisecond to thousands and thousands of resumes.


Chad (24m 51s):

Which is what we saw on Amazon. It's exactly what we saw at Amazon. I mean they were scaling it, but Amazon to their credit, found it and shut it down.


Oana (25m 1s):

But I'm gonna go back to his question. What questions should we ask when we wanna design or buy?


Chad (25m 8s):

Yeah.


Oana (25m 8s):

And I think a problem with what Amazon did was which data was fed into, right? And I think this is also where we need to ask our vendors, providers, is it structured data, unstructured data? Are you working only with our company data? Are you working with a certain nationality? Regionality? There is so much option there that you need to understand before even considering a product. So you can make some decisions where you can say, yes, I can take this risk. Because I think we're still in a phase where everything has a risk. We just need to quantify, is this a risk I can take? And I can work with this vendor on that because you need that continuous maintenance or not. And I'll wait another five years and I'll see what happens.


Chad (25m 48s):

Yeah. Well, so as I have


Joel (25m 49s):

A questionnaire at the EEOC, right? Questions to ask in terms of your recruiting, I'm sure the EU has similar.


Keith (25m 52s):

Yeah, we've put out a lot of guidance on that. But I think to your point, you're making a very good point and I think we need to just set the stage. There's basically two ways AI can discriminate.


Chad (26m 5s):

Only two?


Keith (26m 5s):

Well, two, like, let's just go two major ways. Okay. And that's, you know, either on the data set so you can have a data set that's not diverse. And in the Amazon example, it was a resume set of mainly men. So what happened, it lowered anyone who was a female or had any characteristics related to being a female. And that's because it was, you know, bad data. And that data, even though the data, the characteristics were nEUtral, that was the predominant characteristics. So that's due to the data discrimination, which is what largely everyone is talking about. But the other type of discrimination really occurs from the algorithm itself. So you could have the most neutral, perfect, diverse data set that is representative of your area, that has all the candidates you want.


Keith (26m 48s):

But if you're only then showing those job advertisements to certain people through an algorithm, let's say younger recommender system.


Chad (26m 54s):

Like a recommender system?


Keith (26m 55s):

Or you have a bad HR manager who wants to go in there, like the example I gave you earlier and filter out certain people. That is not based on a bad dataset. That's based on intentional discrimination and that's no different than a human discriminating against by what they see on their resume. So it's very much, it's so important to keep, when you're thinking about AI discrimination, a lot of it is not that much different than human discrimination, right?


Chad (27m 20s):

No, just scales faster. Right?


Oana (27m 21s):

But then is it the system or is still the human? Because I was actually thinking about this, I was putting some people in a LinkedIn project for a recruitment in a certain area, and then I get all these recommended people which are from the same school and I'm like, LinkedIn, I don't need this, I need diversity, please. Because I selected one person and the second maybe tangential.


Chad (27m 44s):

More like Bob.


Oana (27m 45s):

Right? Right.


Chad (27m 47s):

More like Bob.


Oana (27m 47s):

More like Bob. Yeah. Be more like Bob. So in the end when we have this option of people training the algorithm, right, this is the problem as well.


Chad (27m 58s):

Right. Okay. So we've set the table, I think from the standpoint of obviously humans are biased, no question. AI can scale that bias. One of the things that he said though was that right now the employer is on the hook, right? In the EU there's actually guidelines that are being proposed that are being put forth that the vendor will actually be put on the hook as well. So there will be a shared responsibility. So can you talk a little bit about that?


Keith (28m 28s):

Yeah. So this is a really critical issue for companies looking to go all in on AI for, you know, whatever the benefits are, whatever the potential drawbacks. I think with HR technology, we can all agree it is the future. Companies are gonna have to use artificial intelligence in HR, no matter what. But what we're trying to raise awareness of and really the point of this panel is to be competitive, you're gonna have to use some of it, but what do you use it for and what purpose and how do you not discriminate while using it? So you know the regulatory landscape and it is appropriate to start with me on this one because I am a real regulator. It's really.


Chad (29m 6s):

As opposed to a fake regulator like Joel and I.


Keith (29m 9s):

Yeah, exactly.


Chad (29m 9s):

No, I got it.


Keith (29m 10s):

That's what I was hinting at. But you know, in the United States there's exist, like my agency was created outta the civil rights movements in the 1960s. And I've been arguing that our laws from the 1960s still apply to this new technology that's not even that's being developed and will be developed in the future is the same way it applies to decisions made by human, by pen and paper since the 1960s. And a lot of that is just how do we apply these decisions made by algorithms and compare them to the decisions made by humans, which we know how to do. Which a lot of it is results.


Chad (29m 46s):

Which is outcomes.


Keith (29m 48s):

Which is outcomes hiring. We see discrimination. And how did that discrimination occur? Was it based upon, was it it if it's an algorithm involved, was it the bad dataset or was it a discriminatory algorithm? But liability and the discrimination's gonna occur. So what you, what you're seeing across the United States, which we'll talk about on a state and local level, and especially here in the EU, is this rush to make new laws related to artificial intelligence. And probably not a surprise to anybody in this audience. The EU, like they did with GDPR, wants to get ahead of the United States and wants to get ahead of other countries. And they've done that through their proposed EU Artificial Intelligence Act.


Keith (30m 28s):

And in that act, it basically different types of AI use is goes into different risk categories. So they've said that the risk of using AI and employment is in the highest risk category, which subjects it to robust disclosures, auditing, and other requirements. But one of the key differences in the EU if this proposal goes through, is that there will be liability towards the vendor, either those who put it on the marketplace for sale or created and use it themselves. And that's far different than the United States where the vendor does not have liability, the employer has liability. So we are at such a critical time here and this conversation is so relevant and why I'm really excited, I think we're all excited to be here to talk to all of you, especially based in the EU, is that if that goes forward, it completely changes the game when it comes to HR technology.


Chad (31m 18s):

Yeah. A race to regulate is what I just heard. Oana?


Oana (31m 22s):

In the risk of making this EU versus US competition.


Chad (31m 26s):

Let's go. That's fine.


Oana (31m 28s):

You said the EU us? No, I don't think so. I think the EU cares more about its citizens. Can we maybe


Chad (31m 35s):

Agreed. Yeah, agreed. Wait a minute. Wait. Citizens versus the almighty dollar.


Oana (31m 38s):

Yes, but no, I think whatever what Keith said actually is correct and this is happening, it's in discussions technically in the next two years we will know if the law passes or not. But what I'm actually dissatisfied with this law is that it puts everything on the employee or the impacted person. You have to go and say this algorithm from this provider within this company with the lawyer, you know, prepared for all this battle, you need to prove, which I don't think it does anything to be honest. Like right. There will be so little action, like real action that we can interpret this without, like, unless you are, I don't know, you bring some Bloomberg in it or something, right?


Oana (32m 21s):

Like big stuff, but it doesn't really fix the problem of protection and prevention as much.


Keith (32m 26s):

Yeah. But it's put us all on notice. And your point is that the EU even putting a proposal on this really makes us all have this conversation. And it has a different context now of not, you're not just buying a product, but what else do you have to do once you buy the product? And who is going to do that for you? Is it you as the employer? Is it the vendor's responsibility? To your point, is it going to be an employees hiring a lawyer to force those audits to happen? But if this law passes, it adds a whole additional layer onto this conversation, onto the sales and purchase of these products that we're not talking about yet. That doesn't exist right now in the United States, even though the law may from the 1960s may require that audit to occur.


Keith (33m 10s):

So it's just gonna completely change the dynamic around purchasing and developing AI technology.


Joel (33m 15s):

Do you see American companies following suit with European laws in terms of companies being on the hook? Or do you think that they'll continue to get a pass from the US legal system?


Keith (33m 25s):

I think they're gonna follow, obviously in the United States, a lot of these products are being developed, designed there as well. And I think they're gonna keep continue following the US legal standard, but very much like GDPR and how it affected all US companies. I mean as you know, when we're in the United States, we have to click through all those popup GDPR things in the US even though most of us don't even know what it means.


Chad (33m 49s):

Thanks Europe.


Joel (33m 50s):

Accept all.


Keith (33m 51s):

Right. But in all seriousness, I think that when it comes to this, if vendors and employers are gonna have to start doing these audits and paying for the audits here in the EU, the question then purchasing it in the United States, we wanna see those audits. Well you've done an audit and if your audit doesn't show bias, then do it for us here in the United States. So it's certainly will have that effect of adding that additional layer on, if a company has already paid for and done an audit.


Joel (34m 16s):

Can we focus on audits for a second? Yes. What do you mean by an audit? And I think companies being on the hook means you can't just buy an audit from a third party and say that it's legitimate. It becomes a different weight in terms of what an audit is. So define it and what's the future and risks of an audit?


Keith (34m 37s):

Well, from our perspective, from the, at the E E O C and under US law, any types of employment audits, we look at only that individual employer. So any of these generic audits done on, you know, statistical basis of the workforce as a whole does not matter to us. When our federal investigators show up to your business, you know, we're asking only for your data set. We're asking only for the snapshot of your employees. Who applied, who you hired, and that whole chain and looking at those results. So any of these aggregate audits or audits done for other companies or by other vendors mean absolutely nothing to us. And what does that look like? That's the hardest question to answer right now because


Joel (35m 18s):

Everyone's, I wanna, I wanna underscore means absolutely nothing to us, meaning the government agency, that's an important, I think, thing to highlight.


Keith (35m 24s):

We only care about your company and if you come and show us a brochure that it was audited by somewhere else or worked at somebody else's company.


Chad (35m 31s):

Just what it is, is a brochure.


Keith (35m 33s):

We don't, it doesn't fly. And I think that is really, and I'm not saying that, you know, to, to scare everyone. I'm just saying what the, what the law requires and what the reality is. We only look at your employer and everything else that's not going on there. If the product works for somebody else, great, that's great. But we care how it works with your business, with your employees, with your diversity. And that's so critical for everyone to understand. There's not just this, once it's certified, it works.


Joel (36m 1s):

Will there ever be a government agency that audits these AI technologies?


Chad (36m 7s):

Hell yeah. There will be. I mean, okay, so let me give you.


Joel (36m 11s):

Are you sure?


Chad (36m 11s):

Oh yeah, I can guarantee you. Here's why. So back in like 2008, the United States actually shut down a site called America's Job Bank, which was a distribution mechanism for jobs, but it was a part of compliance. So what they did was really the private sector took that on and it became a product. It was productized, although that product had to complete within audit, right? So you had an infrastructure for job distribution and then you actually had auditing information that you had to provide to your clients. I mean to me, from an AI auditing standpoint, every single one of these companies that are out there today should they, they could have a market differentiator if they had a audit and also third party audits to provide to a company which, which again doesn't save you although you go through the mechanism and you actually understand what you're looking for.


Chad (37m 10s):

So that when the guys come in the door looking to do an audit or to check to see if there is discrimination, you at least have information to start to go by. Some type of infrastructure.


Keith (37m 20s):

And I wanna go back to Joel's question. What is an audit? It sounds great, right? Audit, audit, we did an audit, we paid for an audit, we're audited, right? Okay, that's great in accounting, that's great with your finances. Yeah, but what does that mean in employment? And this is what the most difficult part is. And this is where the EU does look to the United States and most regulators around the world in employment. And they look to the E E O C. Because in 1978, that's right, in 1978, the EEOC came out with guidelines in coordination with the Department of Labor and Department of Justice on how to do audit testing in employment. Okay? But that was created in 1978.


Chad (38m 1s):

That was before Oana was even born.


Keith (38m 3s):

And that is the standard around the world. So in New York, for instance, New York City, as you probably a lot of you have heard of, it's the only real auditing law that is about to go into effect in January, 2023. And it's for use of AI systems for recruiting, but then limited to New York City where their jurisdiction is. And they said it's gonna be, if you're using this, you're gonna have to do a yearly bias audit. Well, what does that mean? And everyone's pushing New York, pushing New York, tell us what that audit means because we're gonna have to do it if we're in New York City. So New York just two weeks ago finally came out with what an audit is and you know what they said? Look to the EOCs 1978 guidelines, because nobody has come up with a better system or anything.


Keith (38m 45s):

And that system was really designed for pen and paper Scantron tests, assessment tests, and whether we've all taken one, a lot of companies used to give that, you know, you wanna work here, take this psychological test developed in the 1940s and fifties, you know, the Wonderling tests. There's a lot of different tests and we're gonna see if those tests are biased based upon their results. So that's how those tests were designed. But that's a big overhaul to do that. There's a lot of other things going on. So right now the really the worldwide standard for what is an audit is these 1978 guidelines made for pen and paper tests.


Chad (39m 18s):

Oana how are you auditing bias today without AI or with ai? I mean.


Oana (39m 23s):

That's what I was gonna, is the audit gonna be done by ai?


Keith (39m 27s):

Right. Well there's a lot of companies out there who are, they're now developing ai, right, to audit AI and?


Oana (39m 32s):

Yeah.


Keith (39m 32s):

Sort of a half a joke, but there's really now vendors coming online building AI to see if your AI is biased. And then do we need another AI to check that AI?


Chad (39m 42s):

Startup in the audience.


Keith (39m 43s):

When does it end?


Chad (39m 44s):

There's a startup already in the audience doing that? I'm sure.


Oana (39m 47s):

I don't remember the question, but


Chad (39m 50s):

How do you currently audit? I mean what to what standards? Because there oh yeah, there's still standards, right? And just because it's happening with AI doesn't make the standards any different because it's all about outcomes.


Oana (40m 0s):

But in a way, at least here, we basically if you are a ProducT, dunno HorseFly for example, or Candidate ID or things like that, they need to go and submit their product and say, I am GDPR compliant, I am AI compliant, I am also local regulation compliant because German is different than the UK than France and so on. So in a way, as an organization, usually you're like, are you compliant? Great. Then we do our own compliance with our own legal team. But that's the third party that you say we shouldn't care about, right? So if you work with your PWC or whoever, right? If you can have that, that's good. But a lot of organization do not, especially small and medium, you will not pay PWC X amount to say,


Keith (40m 42s):

Well you're asking companies with smaller budgets to buy these sophisticated products. And then you're asking them to go, like for instance, company like yours, you have a lot of outside auditors, but what about for the smaller companies and how is that dynamic going to work and where is the cost shifting gonna go there? And a lot of these questions aren't answered because no one has forced them to do it yet. The EU law is just a proposal to probably no surprise of anyone in this audience. The state of California is also trying to implement, under their labor department auditing requirements and vendor liability. But we haven't really seen that requirement happen outside of New York.


Keith (41m 23s):

But again, it's just for those New York City residents. So we haven't seen push come to shove on how it's going to then be part of the whole, we're selling you an AI package in HR, does it include this? And where is the cost? So I really think it's gonna be so fascinating coming back to this conference in a couple of years from now.


Chad (41m 42s):

Yeah.


Keith (41m 43s):

To see how that dynamic has changed.


Chad (41m 45s):

Yeah. So, so we're seeing, we're seeing a wave though. The EU we see big, I mean huge states and cities like New York, California, I mean this seems like inevitable.


Keith (41m 53s):

Talking, if it's inevitable. The laws still apply right now and so much of this AI discussion. So.


Chad (41m 58s):

Yes!


Keith (41m 59s):

It's now you can't use AI to discriminate right now. And a lot of people are in the meantime thinking, well we don't have, you know, this New York law is not in effect, California's proposal, the EU. We're okay right now and you're not in no sense. And those requirements to make sure that AI is not discriminating. And there's more than just use in hiring, right? It's being used across the board for managing employees to making pay decisions for employees to even terminating employees. It's being used across the board. And from my perspective, from Washington DC we have to make sure that there's companies in every use of AI in the workplace all are still complying with the laws that have been on the books where started from the 1960s.


Keith (42m 41s):

And whether there's new auditing requirements required by law, it doesn't matter. You still have to this second be making sure they're not discriminating and the second being diligent in who's using the systems and how you're buying them and how you're implementing that. And I think that really gets lost in the conversation when you start talking about the future of AI regulation because there's laws right now that prevent discrimination.


Oana (43m 2s):

What is though being added now, which I think is very interesting from the EEOC. You always add this equal employer statement, right? So you have an obligation, and we do that in Europe as well. Each state has its own kind of guidance. But now at least with this European law, we need to add that we are using product with AI to judge your selection. Like you have to have this statement in your job description throughout your recruitment process so people are aware how they've actually been through a process and what they can action if they want action, anything. And I think that's interesting, but probably in 20 years, 30 years from now on, everybody would've been using it. So the statement is void, right?


Keith (43m 43s):

And I'll quickly take the contrary to that because yes, that you can't discriminate in hiring, everyone needs an equal opportunity. But now with the AI disclosure requirements that New York's having that some of the, you know, the proposals have that, if you're gonna be subject to artificial intelligence, you need all these disclaimers. You have the ability to opt out. Okay? Well now we're going further than how it works now or say if you're dealing with a hiring manager who's racist, a hiring manager who doesn't wanna hire you because you're pregnant or you have a disability. There's no disclaimers now that's saying you're being subject to a somebody in talent acquisition with bias. And you know, you may not get hired, you may think, well I just wasn't qualified. But the real reason you weren't hired is because you're pregnant or you're disabled.


Keith (44m 27s):

And how are we supposed to know that? How do we get into the human mind, the black box of the human mind? But now when it comes to AI, we're holding AI to a much higher standard in a sense where we're making it a lot more confusing because of the black box of AI. Well, I could also argue that AI is more transparent than the human brain, right? So is anyone ever going to admit, I didn't hire you because you're a woman? Is anyone gonna ever gonna say, I'm not hiring you because my bias. But now we're we're saying that the AI will, you know, it needs to prove it on that side. So you know, where are now the different standards of putting AI that is actually potentially more transparent on a different level playing field than a human when the human is the one actually building and designing the AI and I think that's another very complicated conversation.


Joel (45m 13s):

Come back next year when we'll be talking about AI hiring for squirrels and dump trucks in the metaverse. Everybody give it up for Keith and Oana! Guys, particularly Keith, maybe you as well. If they wanted to learn more about these issues, where would you send them? What's a good resource that they can tap into?


Oana (45m 29s):

Half of the vendors here have AI component in their promise, machine learning, deep learning. We will see. Go and check them out. I think, they're gonna teach you what they do and then you can take it from there.


Keith (45m 39s):

If you want to go back and listen to the Chad and Cheese podcast, we did a whole deep dive for an hour.


Joel (45m 45s):

Yeah.


Chad (45m 45s):

About how the technology's being used in different kind of technologies and the regulations. But it's really important for me as a regulator. And from here I'm going to another conference with EU regulators and regulators from around the globe. I think we really want to get this right. I think we're at a really interesting time where, you know, how do we make sure that employers, like you buying these software, have the tools they need, have the questions. How do we make sure that the vendors who normally the EEOC doesn't have jurisdiction over and doesn't deal with, have the tools they make? Cause I truly believe that nobody wants to build a product or sell a product that violates civil rights laws, right? So how do we get to that group as well? So as this conversation continues, I'd really like to hear from all of you, what you would like from all from us, from the regulatory perspective.


Chad (46m 30s):

What questions we can answer, and then what we should, the questions you need to ask a vendor? You know, run them by us. And I think that's how we're gonna be able to make great guidelines as well.


Joel (46m 45s):

I'm Cheese, he's Chad, we are the Chad and Cheese podcast. Thank you everybody. We'll see you soon.


Keith (46m 53s):

Thank you.


Joel (46m 53s):

We out.


Chad (46m 54s):

We out.


OUTRO (46m 54s):

Thank you for listening to, what's it called? The podcast with Chad, the Cheese. Brilliant. They talk about recruiting. They talk about technology, but most of all, they talk about nothing. Just a lot of Shout Outs of people, you don't even know and yet you're listening. It's incredible. And not one word about cheese, not one cheddar, blue, nacho, pepper jack, Swiss. So many cheeses and not one word. So weird. Any hoo be sure to subscribe today on iTunes, Spotify, Google play, or wherever you listen to your podcasts, that way you won't miss an episode. And while you're at it, visit www.chadcheese.com just don't expect to find any recipes for grilled cheese.


OUTRO (47m 33s):

Is so weird. We out.


Comments


bottom of page