Could AI Agents Upend Antitrust?

Bob: So for more than a century, that line has been pretty clear, but what do you think is going to happen when we enter a future with AI? What complications do you see coming?

Arman: Yeah. So one of the key things will be AI will allow most businesses to have effectively employees who are not humans. It’s kind of weird to think about it like that, but a lot of the work will be done by some digital being.

So if there is collusion across different agents, I’m not really sure how that’s going to be contained, how it’s going to be monitored, how it’s going to be enforced.

That new world requires very serious compliance programs, right? From a legal perspective, we need to know what these AI agents are doing, especially when they cross the corporate barrier and start talking to third parties.

Now, there’s no doubt that they will do that. Just like I’m able to send an email to a colleague at another firm, we will have AI agents, tools, whatever you want to call them — employees — that will be interacting with other AI tools and employees and whatever at competing companies.

It’s one thing to bring in a bunch of VPs and tell them about antitrust compliance and what they should and should not do, send around a PowerPoint deck, thou shall not do this, thou shall not do that, and then kind of stay on top of them to make sure they’re not doing it.

It’s quite a different thing to do that in a digital format. It’s going to be a really interesting world.

Bob: And is the issue partly that you may or may not know when the agents are communicating across corporate boundaries, or what makes it so complicated there?

Arman: Well, but that’s one of the issues, right?

How do you monitor the coordination from different companies?

The other issue might be … I could see serious benefits for a bunch of competitors to agree to share information. We do that all the time. It’s called indexing, creating standards, whatever.

We do it with the antitrust blessing, right? Lawyers get involved and say, “Oh yeah, you’re allowed to do that.”

But I could see with agentic AI being so fast and so superior to human intelligence that they would be able to spot these opportunities, right? Opportunities where there are real efficiencies from information exchange.

Forecast our manufacturing more effectively, just to give you an example. But where to draw those lines is not going to be so clear. And more importantly, the speed with which they can do it may very well be beyond our capabilities as human lawyers to get involved in. These are really interesting issues.

I’m not really sure if I were a policymaker which way I would come down on any of this.

But I know these are going to be real issues, and we’re going to struggle with them.

Bob: Yeah, so there’s a big question as to whether we can really know what AI agents are thinking at all. So how would we even know if they’re colluding?

Arman: You’re raising a really interesting question. I mean, antitrust laws presume that humans are trying to maximize profit. I think most policymakers assume that too. Most judges, most courts will assume that in most cases. Now, I think we’re going to assume that about agentic AI also, although it’s not really clear to me how and why we would assume that.

Unlike us mere mortals who are trying to make a buck so we can have a better life — that doesn’t apply to agentic AI, right?

I mean, if they do in fact have agency, what assumptions can we make about their behavior?

Like, how do they operate? What is it that they’re trying to do? It’s an interesting issue, and I’m not 100% convinced that we’ll be able to use the same assumptions moving forward, even for corporate America, because part of their employee base will be replaced by a completely different entity that is literally immortal.

So it’s a completely different thing.

Bob: So question, do you think AI will just force us to come up with new interpretations for the same old laws, or do you think it’s bigger than that? Like antitrust as we know it could cease to exist.

Arman: There is a good argument that the technology, AI technology in particular, and the availability of effectively perfect information to anybody at any point in time might affect how we look at the very foundations of antitrust laws.

Because antitrust laws are designed to protect a competitive landscape in which it is assumed no one has perfect information. That is where the benefit of competition comes from. It is because of the uncertainty that folks compete. It is because I don’t know if I’m going to lose my business to my competitor that I lower my prices.

In the presence of perfect information, we have no way of knowing how a market actually operates. If the availability of AI changes how economists start modeling out what a perfectly competitive market looks like, what oligopolistic markets look like … if that happens, which it probably will happen, I think we’ll need to go back and see if our antitrust laws are still well-equipped to deal with that reality.

Bob: OK, well, let’s bring it back to the present again.

Arman: Let’s. It’s a lot easier for me to talk about the present.

Bob: We’re not at the agentic future yet, but it’s coming fast. What should companies be thinking about when they’re adopting AI and they’re starting to think about agents and this stuff is beginning?

Arman: I think compliance is a key issue. Making sure that you’ve looked at how the antitrust laws might be implicated in your use of agentic AI tools and putting in place some mechanism to be able to monitor that, if not prevent antitrust violations from happening. Some of these steps might not be enough in practice, but attempting to do it is good enough in most cases.

Bob: You’ve given us a lot to think about here, Arman. Thank you so much for joining us.

Arman: Hey, this was a lot of fun. Thanks for having me.

Bob: Arman Oruc is a partner at Goodwin who co-chairs our Antitrust & Competition practice. You can find links to more of his insights in the show notes, and you can check out all of our analysis at goodwinlaw.com. That’s it for our first episode. I hope you enjoyed it.

If you did, here are two things you can do.

One, rate this podcast. We’re brand new, so it really helps other listeners to find us. And two, please make sure you subscribe.

We’ll be back with new episodes every other Monday. Thanks for listening. I’m Bob Mertz, the editorial director at Goodwin, and this is Forces of Law.

This informational piece, which may be considered advertising under the ethical rules of certain jurisdictions, is provided on the understanding that it does not constitute the rendering of legal advice or other professional advice by Goodwin or its lawyers. Prior results do not guarantee similar outcomes.