Workers must have a seat at the AI bargaining table

In this episode of Metro Blueprint, experts explain the importance of including workers’ voices in conversations that shape how AI is used and regulated. Molly Kinder, senior fellow at Brookings Metro, and Christy Hoffman, general secretary of the UNI Global Union, discuss how unions can protect workers from displacement as a result of companies using generative AI.

• Listen to Metro Blueprint on Apple, Spotify, or wherever you like to get podcasts.
• Watch episodes on YouTube.
• Learn about other Brookings podcasts from the Brookings Podcast Network.
• Sign up for the podcasts newsletter for occasional updates on featured episodes and new shows.
• Send feedback email to [email protected].

[“doors opening”; music]

PUENTES: Hi, I’m Rob Puentes, vice president and director of Brookings Metro. I am also the host of Metro Blueprint, a podcast from the Brookings Podcast Network. Every two weeks, a Brookings Metro Scholar and a guest expert discuss ideas and action to create more prosperous, just, and resilient communities in America.

There’s been a lot of talk about how generative artificial intelligence is a threat to white collar jobs. A recent Brookings analysis of OpenAI data found that higher paying fields with advanced degree requirements are most at risk to be impacted by AI. In this episode of Metro Blueprint, Molly Kinder, senior fellow at Brookings Metro, and Christy Hoffman, general secretary of UNI Global Union, discuss what these trends mean for workers and how to ensure that worker voices help shape how AI is used and regulated.

You can learn more about this podcast on our website, Brookings dot edu slash Metro Blueprint. And now, here’s Molly and Christy.

KINDER: Thanks, Rob. I’m Molly Kinder, a senior fellow here at Brookings, and I’m delighted to be in the studio today for a conversation with Christy Hoffman. Christy is the General Secretary of the Uni Global Union, an amazing union representing more than 20 million service workers across the world. Welcome, Christy, we are so delighted to have you.

HOFFMAN: Oh, well thanks Molly, and I really appreciate the invitation to join you today.

KINDER: Great. And the reason why I invited Christy for this conversation is the topic of AI work and workers is not only the subject of my work here at Brookings but is something that it feels like everyone is talking about. Especially the last several weeks we’ve seen dramatic news headlines about everything from the CEO of Anthropic warning of a coming white collar bloodbath, to statements from some of the biggest companies in America citing AI as a reason for either job cuts now or in the future.

What we wanted to do today is have a conversation with Christy about what does this mean for workers. We hear all the time from tech CEOs, from venture capitalists, from economists. It is so important that we make sure that workers themselves are part of this conversation.

And I think I have the greatest job in the world because I get to think about these issues and talk to so many people from MIT professors, to business leaders, to Vatican leaders, and I would say of all my conversation, Christy stands out to me as one of the brightest thinkers on these issues. So forward-leaning and really brings not only deep experience working alongside workers but really understands sort of what the future holds.

So, delighted to have Christy. Can you just give us a little bit, for those who aren’t familiar with UNI, tell us a little bit about your union, who you represent.

[3:11]

HOFFMAN: Thanks. Well, our members are unions, which represent workers in 10 sectors as we define them in our global economy. And the occupations vary wildly. They include tech workers, call center workers, cashiers, bank workers, television writers, even security officers and postal workers. And I’m leaving out many. So wide range.

Many of them affected by AI. We have four regional teams. So we have a global, regional, and sectoral structure through which we carry out our work, which is aimed in part about facing common challenges like AI.

KINDER: That’s great. Now, Christy, in some of our conversations, I was really struck how you started your career on a factory floor as a shop steward and coming out of, I believe remind me your undergrad, you came out of Smith?

HOFFMAN: Smith College.

KINDER: Smith College, and you know, ended up with a very fancy law degree, and here you are several decades later, leading one of the most important unions in the world, frankly. So tell us a little bit about that experience. How has your just incredible career journey, which is unusual, how does that give you perspective into what your members are facing today?

[4:26]

HOFFMAN: Well, it’s true. I did start out as a shop steward in a jet engine factory surrounded by mainly men, and we had a very strong union there. And I wanted to do union work, and I wanted to start with the basics. And it kind of, just bringing this back to technology, I operated a vertical turret lathe. And when the company introduced numerically controlled machines in the very early ‘80s and even late ‘70s, which are kind of precursors to many of the robots you could say that we have today, I was really excited. I don’t have to turn the wheel all day long on my lathe, and I thought, well, this will be enough to take away some of the physical burden of this job.

But at the same time, I had the confidence that our union was strong enough to negotiate protection so that nobody had to really worry about displacement or health and safety hardship.

And so I was on the negotiating team in the early ‘80s. And we negotiated our first new technology clause, a very creative name, and that was over 40 years ago. And I think when I look back at the language that we negotiated back then, a lot of it is really still relevant today. And there’s just such a long and rich history in the United States and in other countries of negotiating around technology very successfully.

KINDER: That’s great, Christy. Well, in my work at Brookings I lead a project here on how generative AI specifically is impacting work and workers. And importantly, what do we do about it? And I have this strap line that for workers AI presents both wonder and worry. And I love your example from your experience because in that example, technology wasn’t necessarily a bad thing. There were some positives that you were able to embrace in part because you had real agency and voice in the workplace. Give our listeners a sense of what you make of this moment today with generative AI. What does this mean to your workforce? You’ve talked about this being something that could either go in the direction of really exacerbating inequality or doing the opposite. Tell us what you make of this and what this technology means for your members.

[6:35]

HOFFMAN: Well, obviously there’s big differences across countries and sectors. But there’s some common themes, and one is that everybody, even those who have less to worry about, frankly, is somewhat fearful about what lies in the future. And that depends on how much confidence they have, that they’re gonna be included in the decisions and where and how AI will be deployed.

Everybody worries that it’s gonna reduce their autonomy or replace them entirely. And they fear losing their livelihoods and also their way of life. And to some extent, we’re assuring people, no, your particular job is really not as much at risk. But of course with all the hype right now and the exaggeration of, well, no, there won’t be any jobs left in five years or 10, people are even more anxious than they normally would be.

But it’s a high topic for workers everywhere. But I think the important thing is that it’s all about confidence. Confidence in that you will have a role. And if you look at some of the unions, for example in Europe where there is what’s called social partnership, it could be an opportunity to get more productivity, better wages, maybe a shorter work week.

When we look at Germany where the Works Councils have a lot of power in negotiating around technology and and really shaping the use of technology, they’re not as worried about job displacement. Or Japan is not as worried about job displacement. And Northern Europe is not. Then, you know, when you look in Asia and Africa, I’m really surprised how much tech optimism there is there, but in part because it hasn’t been deployed as much as it has been in the U.S. for sure.

And then of course, certain sectors are much more vulnerable. Customer service workers across the world, we’re already hearing stories of reduction in staff, and that’s been a really big event that’s unfolding even in the past six months. Some of the call center workers, they like having the AI coaching on how to answer the questions, but they’re not so happy that they have to work harder without additional pay. They’re not so happy to see some of the jobs disappear. And if you look at banking workers and media workers, they are very, very nervous. And those two sectors are the most deeply involved in collective bargaining.

So there’s a wide range, but the key point that we try to drive home is that we have to be on the front foot. We have to be demanding to be at the bargaining table with these employers to get some commitments about notice, about consultation, and about other things that are relevant to the jobs involved.

KINDER: Well, Christy, so many things you said hit home to me first the point that you have a very diverse membership, and many of the sectors where your workers are employed are really according to our data at Brookings, where we have a data set from OpenAI, when you look at what are the types of tasks and occupations and sectors that are the most exposed to this technology, meaning where you could use gen AI technologies. You’re in some of the most exposed. I mean, finance is really out in front, the media industry, customer service.

And media is a good example. It’s not only just how the media industry adopts this technology, but it’s how much the whole sector is changing as a result of gen AI in terms of Google’s search is summarizing results. So you’re really in the crosshairs, I think of a lot of change.

And I really want our listeners to hear you drill down a little bit to a very important point you made, that the exact same sectors in America, the workers in those sectors here might actually feel more anxiety than workers in those exact same jobs and exact same sectors in other countries because of just how different our policy environment is, our institutional environments. Germany, I think is a great example. For those who might not be familiar with some of these German models, can you share a little bit of what that looks like? And obviously you know America well and contrast that a little bit to our institutional setup.

[10:40]

HOFFMAN: Well, first thing that’s really interesting that I’m just kind of getting my head around, and I do this for a long time, but I didn’t realize that when Works Councils were established a very, very long time ago, well before World War II, they were built in with the requirement that the employer should consult with the workers, with the Works Council about technological change. And this was a key element. And so they have a big role in it.

But it’s interesting because when I was at the World Economic Forum last year, and I was asked, well, what would we want? And we said, well, one thing is we want notice before an employer would intend to introduce AI, the workers should have notice and an opportunity to do an assessment and so on. And they said, well, how much notice? And I said, well, let’s say three months. And the employer representative on the panel from Germany, and I won’t say what company, she said, well, we could never get away with three months’ notice. We have to give much longer notice to our Works Council. And you know, she said we would never even think about it, giving three months’ notice.

But I think that’s just a reflection, that there’s an entitlement to resources to help understand the technology. There’s transparency requirements. And the Works Council does have the power ultimately to block the technology from being introduced. But rarely does that ever happen, because there’s an invested interest in seeing the successful deployment of technology.

And I think that’s kind of a myth that a lot of U.S. employers have is that, well, if they had a union, the union would just make it impossible to operate. Whereas the opposite is true, is that where you have a Works Council in Germany, and you normally also would have a union as well as part of that structural arrangement, they’re eager to make it work. They just want to make sure that people aren’t harmed in the process, but they’re very eager to make it work. And I think they’re doing great work in terms of bringing everybody along, making sure the right people get transferred to new jobs and or upskilled for new jobs.

And I think they have a very good model. But it’s built into their model that this is what you do when you’re changing the way work is organized.

KINDER: So, Christy, given that we’re having a real moment in the U.S. where all of a sudden the topic of AI and work is starting to capture political attention. So what’s really striking is just in the last few weeks, you’ve had both left and right, from Bernie Sanders on the left to Steve Bannon on the right, Pete Buttigieg just raised his hand and said this is an issue he’s gonna pay more attention to, Senator Chris Murphy. This is starting to appear mainstream, left, right, and center. Yet in the U.S. our system could not be more different from Europe and there is a lot of anxiety. We have terribly low union density, not only nationally, but particularly in the sectors where AI is gonna be the most disruptive.

This is something Mark Muro and Xav Briggs and I have coined the Great Mismatch: 95% plus of American workers who are in the types of jobs where AI is gonna change the most don’t have union membership. There’s no collective bargaining. We don’t have this concept of in workplaces a Works Council where workers have to sit together with their employers and talk through these things and work together to figure out, yes, we need to adopt AI to serve business goals, but let’s bring workers along with that. We don’t have the same investments and training. We have a much weaker safety net.

So what would be your message to policymakers in America who are starting to really take this seriously? And understand as a country where we’re going all in on AI. What does it look like to go all in on American workers?

[14:20]

HOFFMAN: First of all, I’m a trade unionist and I will not ever be persuaded that we should have alternatives to unions to take up this issue. I think that there’s some discussion. We should have tech committees or just similar to health and safety committees, and a lot of that’s out there. I don’t really think that we should have as many obstacles to bargaining in the United States as we do, and there’s a lot of ways to change that. It does take political will, which is not overflowing that bucket right now.

But I think that the tech companies themselves have a responsibility to promote bargaining as a social good, because it’s a bit of an anomaly to say there’s so much progress that’s gonna be possible from this, and the world’s gonna be a better place, and we’re all gonna have so much stuff and at the same time say, but workers should not have unions, which too many companies are taking that line.

[15:26]

But beyond that, there’s a lot of other things, legislative things that could improve things. I mean, for example, we had the WARN Act for layoffs in the past. I don’t even know if it’s still in effect. We can have requirements that companies be transparent about their intentions to implement AI, that they give notice to workers that there are some possibilities of challenging it or raising objections. So that’s one thing. Getting notice is very important. We know that that has made a difference in many other cases and kinds of issues.

And the second thing is, and this is personal to me, but I just feel like it should not be so cheap in the U.S. to get rid of workers. I think we need to have some costs attached to it. European employers are much less likely to make their workers redundant because they have to pay severance pay. They have to give notice. They usually are required to consult. And I do think that some kind of severance pay issue should be thrown in there.

I would point out the AFL-CIO has a good list of legislative proposals. But one of the things that I’d like to throw out there is that there should be incentives for companies to retain their workforce in times of disruption. Maybe tax incentives for employers who want to train people for transition to other roles on the job. Maybe shorter work weeks. All of those things need to be on the table because it really should be a priority to keep people at work.

Definitely there can be legislation, and I know there’s a lot out there already to limit surveillance, and a lot of the gen AI is coupled with surveillance. So the surveillance is, like, let’s just say at a call center, you’re getting constantly surveilled at the same time you’re getting answers on how to answer the calls or feedback. I think that there needs to be some limits on that because that’s proven to be so hard on mental health and denying privacy.

And there are other ideas. And I think that we have to obviously think about the safety net. And training. But I’m also skeptical of training for training’s sake. I think it has to be embedded. Labor management training is certainly a key priority, but also I’d really like to see people train for jobs that do exist on the job. And I’m nervous that too often employers will think, well, rather than retrain my existing staff, I just can go hire some new people and they’ll be cheaper and maybe better because they’re digital natives or whatever the reason is offered.

So I think the question of on the job training for new jobs is important. And if you notice the recent announcement among Amazon workers that there would be a huge disruption. There’d be a lot of new jobs and a lot of jobs lost. And they weren’t sure about the net. But there’s no guarantee that the ones that are lost are gonna be moved over to the ones that are gained. And I think that’s something we have to keep in mind.

KINDER: Great. Thanks, Christy. That’s a great robust list.

I’m curious, you’ve used some language in the past about a just transition, which I know we often think of in reference to climate change. Studying this technology, it seems very clear that in the years ahead, workplaces are going to change, different skills are gonna be in demand, jobs are gonna look different. What does it look like to do this in such a way that acknowledges that AI is a tool employers are going to use, that all of us in the workforce at some point, AI is gonna be part of our working lives?

But how do we do it in an ethical and a just way that brings workers along? I’ve had conversations with unions in America where some of them are really leaning in. If you look at education unions, it’s actually teachers, not students who are leading the charge with AI, because it’s helping them do their job. And so the question is, how do we make sure that teachers are driving some of the decisions, they can adopt these tools and that the teaching jobs of the future retain that human touch and their dignified jobs?

You can imagine that in a bank and in a hospital and in a call center, these jobs are going to change. What does it look like for this to be something that workers are able to transition to the way that work is going to change in a positive way? What does that take and what sort of institutional investments need to be made to enable that transition?

[19:51]

HOFFMAN: I think you raised a really excellent point, and I think that the teachers are doing great work. And there are so many examples of where workers are like, we want to use this technology, and we could see a lot of opportunities and ways that we could use it better than even what our employer is suggesting, but we need to be part of that discussion. And we can’t be part of that discussion if we’re just worried about our own situation.

I think that there’s so many studies that have been done that show that technology is way more successful if it’s done together with the workers who are gonna to be using it and they have good ideas. They can maintain their humanity through the process, but at the same time maximize some of the productivity enhancements or see new opportunities.

So it comes back to the question of getting on top of it before the employer wants to implement technology. Starting that conversation, identifying ways that it could be used for the benefits of everybody. And working on it together. I’ve met with a Finnish union in banking who described in detail a year long process of bringing everyone in the bank around the idea of using AI, and everyone at the end was super happy with the results, even though it meant a lot of them had dramatic changes to their jobs. I was really impressed. It took a while. I mean, they did trainings, they did workshops, they tried to identify opportunities and anxieties.

So again, it’s a question of trust. And trusting that workers actually want normally the best for the whole community in which they work. And also trust that if you’re working harder, you’ll see some benefits from that. And I do think there’s a trust deficit on both sides often in the United States, but I think that can happen.

[21:40]

And when we talk about just transition, you know, when it comes to climate, we’re often talking about making sure that the people who are replaced because they’re a coal miner or in some kind of energy job, they’ll find another job. And that hasn’t been very easy for those coal miners in West Virginia, for example, or, or in Australia for that matter.

I think when we talk about just transition for AI, it’s a similar concept that if your job is gonna be eliminated, that you will be woven into another position or there will be some other opportunity for you, that you won’t be one of the ones that are just left behind. And people say, oh, well the net is a job growth, but there’s 8 million people, or how many million people who won’t be part of that growth and they will be part of the loss.

When we heard at the WEF [World Economic Forum] that 22% of all jobs are gonna be disrupted by 2030, 14% are new jobs and 8% of the ones eliminated, but we don’t know if those 8% of those people are gonna actually take on those new jobs. And I think again, that’s part of just transition is making sure that some people don’t suffer disproportionately so that others can progress.

But there’s many examples of employers who really, really are working to bring their workforce along and be supportive of the journey and try to make it a success.

And I, I would add that from my time in manufacturing, that in manufacturing they know that putting a new machine on the factory floor without really getting the support of the workers, it’s really hard to make that work. You really do need to bring people along and see them as partners and not as people who just want to make trouble or slow things down.

KINDER: That’s great. Well, Christy, my last question for you is you started your career on the factory floor. So if you were to give some advice to a similar young 25-year-old shop steward who is in her start of her career and is looking out and seeing that AI is going to be a truly disruptive force, what advice would you have for her?

[23:53]

HOFFMAN: I would just say, you can do it. And I say this to our youth all the time. I mean, unions have been successfully negotiating around this issue for a really long time, and we all know that when unions and workers are involved from the start, it’s better for everyone. And you don’t have to understand how a large language model works. You just have to rely on your instincts about what’s fair.

And I think that’s the key message is that nobody wants to be monitored the whole day. If they work harder, they want to get paid more. All of these little things are, any shop steward will understand the issues on the table and be able to make a good fight about fixing those issues.

And none of this is inevitable. And with your voice on the job, you can make a difference. So that would be my message.

KINDER: Great. Well, Christy, thank you so much. This has been a wonderful conversation. I really appreciate your leadership on the world stage in all sorts of international fora to really bring the voices of workers into the conversation about what our AI future should look like.

We do lots of research on this topic at Brookings. We publish regularly on AI work and workers.

[music]

You can check our website Brookings dot edu to find out more. And thanks for following us for this conversation.

PUENTES: Metro Blueprint is a production of the Brookings Podcast Network, found online at Brookings dot edu slash podcasts. Thank you for listening.

My thanks also to all the scholars and guests for sharing their insights and expertise, and to the team at Brookings that makes this podcast possible including Fred Dews, producer; Erin Raftery, associate producer; Gastón Reboredo, audio engineer; Daniel Morales, video editor; Leigh Balon, Brookings Metro’s director of communications; Carie Muscatello, our graphic design and web publishing manager; as well as our government affairs and promotion colleagues in the Office of Communications. Katie Merris designed the beautiful logo, and Phoebe Copeland recorded the doors audio.

For more information about us, please visit Brookings dot edu slash Metro.

I’m Robert Puentes.

[“doors closing”]