AI-Driven Enforcement in the E-Commerce Marketplace

Justin Pierce: 00:13

If you're a business leader or general counsel, you already know AI isn't just another tech trend. It's the next frontier, reimagining and reshaping how companies operate and compete. With that transformation comes complexity around intellectual property, data rights, regulatory oversight, litigation exposure, and brand integrity. This podcast is designed to give you clarity. I'm Justin Pierce, cochair of Venable's intellectual property division.

Justin Pierce: 00:43

In this season, I'll talk with colleagues and industry leaders about how AI is reshaping business, ecommerce platforms, toy companies, industrial automation, luxury brands, beyond, and the legal strategies companies need to protect innovation while moving fast. Our goal is simple, to help you turn our legal insight into your competitive advantage. Welcome to season two of AI and IP, the Legal Frontier. This week, we're talking about AI and ecommerce platforms. Hi.

Justin Pierce: 01:23

I'm Justin Pierce, a partner at Venable and co chair of the firm's intellectual property division. As e commerce continues to expand globally, so do the risks posed by counterfeiters and sophisticated fraud networks. But ecommerce platforms are not standing still. From AI powered proactive monitoring to cross border criminal enforcement, some platforms are deploying advanced tools and global partnerships to disrupt counterfeit supply chains at scale. In this episode, we sit down with the Director of Amazon's Counterfeit Crimes Unit, Kebharu Smith, who will explore how AI driven enforcement, technology, litigation, and international collaboration are reshaping brand protection and anti counterfeiting efforts in today's digital marketplace.

Justin Pierce: 02:12

Kebharu, welcome.

Kebharu Smith: 02:14

Thanks for having me, Justin. Looking forward to it.

Justin Pierce: 02:17

Kebharu, how did you come to the law? Before we get into your position and formal role, how'd you get into the area of anti counterfeiting and brand protection?

Kebharu Smith: 02:26

Yeah. I'll start with my background in law. I went to Thurgood Marshall School of Law in Houston. It was a childhood dream for me to practice law. I wanted to be a litigator, a trial attorney my whole life.

Kebharu Smith: 02:40

It was during my time in law school that a friend of mine who said, Kibaru, if you wanna try cases, the best path for you to try cases is to go to a local district attorney's office. And though that wasn't my passion at the time, especially in light of my feelings and passions in other areas, I did start my career in prosecution in Fort Bend County, Texas. And it was during that time I tried many cases from driving while intoxicated cases to assault cases, and ultimately, you know, murders and capital murders when I was in Houston, Harris County DA's office. Tried dozens, if not close to a 100 cases on the state and federal side. So trying cases and my passion for prosecution, that was my desire all throughout law school.

Kebharu Smith: 03:29

It was in 2008 that I joined the US attorney's office in Houston where I started prosecuting federal cases from bank robberies, fraud cases. And within a couple years of being in the office, I received my first counterfeit case, and it involved counterfeit Viagra. The story in that case, the background was that a customs officer doing their work at the border protection facility in San Francisco received a package of goods. And when they opened the package, inspecting to protect consumers and citizens in The United States, they saw a box of loose pills.

Kebharu Smith: 04:05

They had an idea that this looked like a counterfeit Viagra pill, and they compared it to the real pill that they received from the brand who partners with Customs and Border Protection. So clearly, had an address that this package was going to in Houston. They worked with US Homeland Security in Houston and did an undercover delivery to a residence in Houston with the undercover agent acting as a postal inspector. And when the product was delivered to the house, the undercover asked the resident, hey, we received this package of loose pills or these yours.

Kebharu Smith: 04:38

They admitted that they were theirs. That person was placed under arrest for receiving and or trafficking counterfeit goods. They then ended up cooperating in an investigation that led us back to Chicago, and it was a global investigation that involved counterfeit Viagra pills. That was my very first anti counterfeiting case, but it was that work was steeped in a career of criminal prosecution, prosecuting a range of cases to include crimes like organized retail crime that often mixes with counterfeit cases. So that was my very first case at the US attorney's office in Houston.

Kebharu Smith: 05:12

After spending several years in that office, I then joined the Department of Justice's Computer Crime and Intellectual Property Section, which is based in Washington DC. And there, my focus was to go after a range of intellectual property criminals from trade secret thieves, intellectual property infringers, and copyright pirates. And there, I prosecuted a host of IP crimes. My last IP crime that I prosecuted was a counterfeit cancer medicine case. Similar to the first case, it was counterfeit pharmaceuticals.

Kebharu Smith: 05:44

In this investigation, it reached back to Houston again. But in this instance, Ukrainian nationals were selling counterfeit cancer medicine that was being sold across the globe, and we were able to convince them to come to The United States and prosecuted them for trafficking and counterfeit goods. But in between that, I prosecuted a host of IP crime cases, and I've had a passion for it.

Justin Pierce: 06:04

That's a great background, and it really shows from start to even finish how global anti counterfeiting can be, and how much the job in terms of what you've done really requires the ability to look at all aspects of the introduction of the, in this case, counterfeit pharmaceuticals to the market, all the way through the investigation to the prosecution, and ultimately, you know, meting out justice to those who are involved in that.

Kebharu Smith: 06:31

You're exactly right. It's we we like to say that two things. One, that counterfeiting is one of the oldest crimes on the books. It's been around. We've seen it in different variations from being sold in brick and mortar to flea markets or just however it gets sold, but it's also a crime that is borderless.

Kebharu Smith: 06:48

Bad actors are not respecters of borders. They're not respecters of laws, and so they're trying to move these counterfeit and fake products just to make a buck. And so when I was a prosecutor, now that I am at Amazon, we have to find ways to meet these bad actors where they are by partnering with law enforcement and brands all over the globe to make sure that we're holding them accountable, and that we're working up the counterfeit supply chain. But I'll use that moment to segue into how I got into Amazon. It was during my time at CSIPs that I received an outreach, if you will, from a recruiter on LinkedIn, which is a good message to keep your LinkedIn profile, up to date.

Justin Pierce: 07:30

Always good advice.

Kebharu Smith: 07:33

And in that instance, I was prepping for a trade secret theft trial in Utah. And I saw the note, it was, hey, Kibara. Are you interested in talking about a potential role at Amazon? And in my mind, I I laughed it off because I've been committed to being a prosecutor. I've been one for nineteen years, and I was also in the midst of prepping for trial.

Kebharu Smith: 07:52

And as an attorney, as at Venable, you understand that when you're in trial prep mode, nothing else matters. And so I said I can talk briefly about this potential role, and I spoke with the recruiter. I let him know I was prepping for trial, and he said there is a team that's going to be started in Washington DC that's going to do effectively what you're doing at DOJ, and that's going after IP infringers. And it piqued my interest, but again, I was in trial mode. Long story short, I didn't get that role, but a colleague from DOJ got the role and invited me to join her on this team.

Kebharu Smith: 08:26

I was a day one member of our counterfeit crimes unit, started in June 2020, and the mission is pretty direct. And, it's a mission that we've lived by since our team has been around for the past six years, and that is to disrupt and dismantle counterfeit organizations and bad actors wherever they are in the world if there's any nexus or touchpoint with the Amazon store. And so what that means is that when we see bad actors are attempting to sell their counterfeit wares or products in the Amazon store, We partner with brands and law enforcement and other stakeholders to hold them accountable. And that's what we've been able to do, and I look forward to talking to you about what we've done over these past five and a half years.

Justin Pierce: 09:05

Thanks. That's a great background and a great, segue as we talk about that mission. It's interesting given that you're trying to meet the counterfeiters wherever they are and whatever way they're having an impact on the Amazon store and experience. That naturally means that your enforcement actions, your investigations, maybe even actions behind the scenes are gonna be pretty expansive. What are some of the examples or maybe impact examples you have of just how far that reach has gone in terms of steps you, your team at CCU have taken to keep consumers safe on the platform or to go out and stop infringers that you discover are involved in bad activity?

Kebharu Smith: 09:45

Alright. Great question. And let me just start with Amazon. Our mission is to be the world's most customer centric store. We're customer obsessed, and we want to ensure, and we do ensure that when customers come into our store, that they can reliably purchase products that they want, that are authentic, that are free from infringement, and that we're stopping bad actors from even operating in our store.

Kebharu Smith: 10:08

And what are we doing to do that? We have teams of thousands who are building tools such as AI tools and large language models that are protecting the store and customers and brands before products are ever even listed. It's the work that's happening behind the scenes that customers and consumers may not see, but it's happening to the tune of billions of scans on our store, thousands, hundreds of thousands of takedowns before an infringing product is ever listed. That type of work that is built or teams that are built to do this work and algorithms and the most advanced technology. Visual learning where we're looking at trademarks and comparing them to what may be a suspect trademark or suspicious mark in the store that's taken down.

Kebharu Smith: 10:54

And when this happens, we're also taking down other related accounts if it's related by IP address or banking information. So much of this work happens behind the scenes. We talk about that. We describe that as our proactive controls. So that's one pillar of what we're doing before the counterfeit crimes unit even gets involved.

Kebharu Smith: 11:15

The other piece of that proactive control is the seller verification, making sure that we have integrity of who our sellers are and making sure that we're verifying who these sellers are. Another angle that we're doing or another pillar is to provide brands the autonomy and the ability to partner with us to go after suspect listings. We have tools such as our Project Zero, which gives brands the ability to take down listings of suspect products. We also have tools such as Transparency, which allows a brand to affix their serialization onto their products so that when we receive them in our fulfillment centers or when customers receive them, they can scan that for authenticity. And we've used this tool for over 2,500,000,000 products to show that those products are authentic.

Kebharu Smith: 12:02

Over 88,000 brands are using this transparency tool. These are some of the tools that we're using that brands can use to help protect themselves in Amazon store. But in the rare instance that a counterfeit makes it into our store, teams like our counterfeit crimes unit goes after these bad actors at scale. We partner with brands and with law enforcement, and we share data in real time when we identify a bad actor. And what does that data look like?

Kebharu Smith: 12:28

It could include information such as an IP address because it's critical to know who's using a phone or a device or a laptop, and that IP address can help us identify that bad actor. We share data such as banking information depending on the country we're working in. We share physical addresses so that law enforcement can move quickly, and we share counterfeit sales information. And when you share this information with brands and law enforcement, it helps them move quickly to go after these bad actors. And we've heard in instances that law enforcement has been able to accelerate their investigations by six to nine months based on the data that we share with them in real time.

Kebharu Smith: 13:07

The last piece of these pillars is the fourth pillar, which is consumer education. We educate consumers about these, you know, the ills of counterfeits and how we're going after these bad actors through public service announcements and partnerships with organizations like ENTA, and remind them that counterfeiting is not a victimless crime that is connected to human trafficking, terrorism financing, and it impacts innovators and creators.

Justin Pierce: 13:33

Thanks. That's pretty thorough. A number of those pillars I can imagine have become more robust since you joined CCU since June 2020. What role would AI play? You've mentioned it some and implied it in some of the different types of tools, but I'm curious, maybe some specific examples of where AI technology has enhanced your capability to either action on some of those tools, improve outcomes, and just do a better job of the proactive management, proactive monitoring that you're doing on the on the platform.

Kebharu Smith: 14:03

Yeah. At the outset, when I mentioned the proactive controls, AI is foundational to that. Those large language models, the vision learning that is looking at what's being posted in the store by sellers, what brands provide us, trademarks and word marks to compare the authentic marks and word marks to what bad actor sellers may attempt to put or place in the store. That's vision learning. That's large language models.

Kebharu Smith: 14:30

That's AI that's being used to take down suspect listings. It's being used to scan the store on a daily basis to make sure that we are proactive. It's, again, that work that's happening behind the scenes that consumers may not fully understand or may not see, but it's the proactive controls. It's that team of thousands and an investment of over $1,000,000,000 that we make to make sure that we're keeping our consumers, customers, and brands safe from the scourge of counterfeits, and we're really excited about what we're doing. Within CCU, our counterfeit crimes unit, we're building our own bespoke tools that we're using, AI models that are helping us, you know, be more efficient in our investigations and speed up and take the fight to counterfeiters more aggressively and quicker.

Kebharu Smith: 15:18

And those tools are tools that are built by team members and using data from across the store. As you can imagine, Amazon has a vast amount of data that we use to ensure that we're moving quickly to go after these bad actors.

Justin Pierce: 15:31

That's great to hear. I think in terms of just what you've done and based on my knowledge of the e commerce industry writ large, this has to be one of the better examples out there of proactive use of AI. Even in what you described, you hit computer vision, large language models, machine learning, that's really a multimodal application of AI. And I think that's sort of unique in terms of its comprehensiveness. Exactly.

Justin Pierce: 15:57

So it's great to hear.

Kebharu Smith: 15:58

We're really excited about it. And we use the information that we learn from bad actors to retrain. And so it's not a static approach. It's we're dynamic in taking information that we learn to then retrain and make sure that we're scaling our efforts. The other thing that's critical about what we've been able to do is that I I mentioned at the outset that this is a global fight.

Kebharu Smith: 16:18

And so we, at CCU, when we in the rare instance that a counterfeit product is sold in our store, we're partnering with law enforcement all over the globe. We have two primary ways that we go after these bad actors. One through civil suits, where CCU will cover the cost of a civil lawsuit when we go after a bad actor. But in the criminal landscape, we share data with law enforcement all over the world. And in 2024, we had over 60 raid actions in China that resulted in the detention or arrest of a 100 manufacturers and counterfeiters in China.

Kebharu Smith: 16:55

So what that's doing is it's not only protecting Amazon customers, but consumers broadly. Because when we're working our way up the counterfeit supply chain, we are stopping those products before they leave these locations. We're also enforcing and supporting enforcements in places like Turkey and places in Europe like Germany. So our mission is global, and we're going after the source. We're going after the counterfeit supply chain with the mission of disrupting and dismantling these counterfeit organizations.

Justin Pierce: 17:26

Good stuff. Now if we flip it around, given what you do, how much have you seen counterfeiters and bad actors using AI to do what they do?

Kebharu Smith: 17:35

There are some instances where we have seen bad actors using tools, LLMs, and AIs to attempt to commit fraud and promote their schemes. Our team also CCU also pursues individuals who try to send or support or submit fake reviews. Right. That's part of our mission. And we've seen bad actors use AI models in an attempt to promote fake reviews in the Amazon store.

Kebharu Smith: 18:05

Let me just draw back for one second. When people come into Amazon, not only do they want to know they're buying reliable authentic products, but they also wanna know whether when they're looking at a seller's page that they can rely on that review or reviews that may be listed. But naturally, bad actors are always trying to find a way to get around or cut a corner, and we have identified schemes where bad actors will pay people to promote a product with a fake review. Our team goes after those bad actors. And we've seen in some instances where bad actors are using large language models and AI to submit fake reviews in the Amazon store.

Kebharu Smith: 18:45

We go after those bad actors globally as well. Right? And so we're seeing AI being used for those types of fraud as well. The the the other piece is that we all know that we're still early in this AI journey. Right.

Kebharu Smith: 18:59

Most of us are learning by the day and incorporating these tools into our daily lives and our professional lives. So we're at the tip of this. We're just now scratching the surface. So we're always looking to see what new ways bad actors are using these tools, but also how we can use them to stay a step ahead of them and meet these bad actors where they are using innovation and AI tools and large language models.

Justin Pierce: 19:24

One follow-up question on that. Earlier, you mentioned some of the civil enforcement and criminal enforcement types of prongs that you take with your work. Given your extensive history at DOJ as a federal prosecutor, can you give us some examples of the types of enforcement litigation, civil and criminal, that you've taken a part in or assisted in?

Kebharu Smith: 19:44

Certainly. And I think it's also important to mention that CCU, with that question, is a global team. We have team members from Seattle to Shanghai and many places in between. I sit in Washington DC. We have team members in London, Milan, Singapore, Los Angeles.

Kebharu Smith: 20:01

We also have teams of former IP litigators and other former federal prosecutors and data scientists and former law enforcement. And with those skill sets, that multidisciplinary approach, we're bringing all tools to the table. So in The United States, when we pursue a bad actor, civilly, we file the lion's share of the majority of our lawsuits in the Western District Of Washington.

Justin Pierce: 20:25

Okay.

Kebharu Smith: 20:26

And we are using claims like the Lanham Act and breach of contract because when a bad actor comes into our store and sells counterfeits, they have violated the terms of service, and we then can pursue them in partnership with the brands through a civil action. And so we filed a host of civil actions in the Western District with brands from small and medium women owned companies like JL Childress. We partnered in two lawsuits with YETI. We partnered with several other brands. When we filed these civil lawsuits, we use the discovery to also work our way up the counterfeit supply chain.

Kebharu Smith: 21:00

So that's the civil aspect is a part of our strategy to make sure that we are getting data on where these bad actors are located, where they're receiving this money, and where the money may exist so we can pursue those illicit proceeds, and we'll also package that information in our criminal referrals. This can also start from a notice of infringement filed by a brand, where a brand may make a purchase and they may notify us this was a counterfeit product.

Kebharu Smith: 21:27

We use the data and the takedown information from within our store to use that information and referrals and turn it over to law enforcement. And those notices of infringement in some instances have led to enforcement actions in China, where not only the brand that notified us of the counterfeit, but other brands have been favorably enforced where products have been seized. So it doesn't just help Amazon consumers or customers, it helps consumers and brands broadly because notices and the takedown information can help us work our way up the counter for supply chain.

Kebharu Smith: 21:59

We've seen that with brands like Pandora, where it was a multiyear investigation where we partnered with Pandora, and we were able to enforce against bad actors in China where bad actors were held accountable and sentenced to jail time for selling counterfeits in the Amazon store or attempting to do so.

Justin Pierce: 22:14

It really highlights the theme. It's come up in a few of your answers already today about how important it is to collect and then share information to get better at any counterfeiting brand protection. Protecting the platform, and then you've got this other benefit we've talked about a few times, where it actually helps out consumers in general or other brands if you're information sharing. How does that work outside of consumers and brands, but let's say on the criminal side, what's the information sharing experience like when it comes to a customs enforcement or working with law enforcement in this area?

Kebharu Smith: 22:47

One of the reasons that this team is I believe, appropriately staffed with former law enforcement and prosecutors is that we understand the nature of criminal investigations. There may be a time where we share data with law enforcement, and law enforcement can't tell us as much as we told them. So we understand that it's not necessarily reciprocal. But the point is to help them move quicker and to help them enforce against these bad actors, and that's what we've been able to do. We've had enforcements all over The United States.

Kebharu Smith: 23:12

We've had enforcements in Europe, and as I mentioned, in China. And we're also expanding now into India, and we're really excited about taking our team to India to go after bad actors there. But the mission is to either hold them accountable in the courts and through criminal enforcement or to receive judgments and claw back those illicit proceeds, but to also keep them out of the Amazon store from selling their counterfeit products in the Amazon store. That is our mission. And then lastly, to work our way up the counterfeit supply chain.

Justin Pierce: 23:39

Excellent. So given the conversation today, we've covered a lot of things global, we've touched on AI, anti counterfeiting, the great work of CCU. And as we bring this to a close, are there any lessons learned or takeaways that you want the audience to remember from our discussion today?

Kebharu Smith: 23:58

Again, thank you for having me. And I will say that fraud prevention, anti counterfeiting efforts, and customer protection in the ecommerce age is constantly changing, especially here at Amazon. We are constantly investing to protect our store and consumers and using cutting edge technology to protect our consumers and brands and customers from fraud, abuse, and counterfeiting. Lastly, we're taking the fight to these bad actors in an aggressive way by pursuing these bad actors wherever they are in the world to hold them accountable in criminal courts and to claw back any illicit proceeds that they may have gained.

Justin Pierce: 24:35

Okay. That's all we have time for today. I want to thank Kebharu Smith for helping us better understand how Amazon is using AI, data, global partnerships, and both civil and criminal enforcement to disrupt counterfeit and fraud networks and protect consumers in the ecommerce Marketplace. You can read more about how Venable is helping businesses navigate the AI frontier by visiting venable.com/ai. Please join us next week when I talk to my colleague, venerable partner, Meaghan Kent, and Ethan Cohen, global head of privacy and AI compliance at Gap about how to build principled, trust centered AI programs that move at the speed of technology without sacrificing rigor.

Justin Pierce: 25:25

I'm Justin Pierce. Thanks for listening to AI and IP, the Legal Frontier.