Fighting Surveillance Tech with Trademark Transparency: Workshop with Decoding Stigma

Decoding Stigma and Hacking//Hustling present a workshop with Gabriella Garcia, Kendra Albert, and Amanda Levendowski.

Sex work stigma has been used to justify research funding for surveillance technology that profiles and further harms at-risk communities. From non-consensual image and data scraping to algorithmically-informed deplatforming, the sex worker is both indispensable yet disposable for those behind the development of these technologies.

Amanda will teach you how to access the Trademark Electronic Search System (TESS) and use trademark disclosures to take action against surveillance technology in development.

Let’s get started.

Gabriella Garcia: So I am Gabriella. I am here on behalf of Decoding Stigma along with a couple of other comrades. I am on the unceded land of Haudenosaunee peoples. I urge you to research the communities that you’re located on. And this acknowledgement demonstrates a commitment to the beginning process of working to dismantle the ongoing legacies of settler colonialism.

But beyond just acknowledging this is extraordinarily relevant to our conversation today. The ongoing genocide of indigenous bodies is a product of state surveillance that protects extractive capitalism and incarcerates and disappears some of our most marginalized comrades.

I must also pause for a moment to highlight that today is trans day of remembrance. In the past 12 months, 350 trans and gender diverse people were reported killed globally and sex workers do make up the majority, about 62% of victims. Enforcement of gender conformity is a surveillance practice.

Bathroom bills are a carceral technology that endangered trans bodies. Walking while trans laws that disproportionately target trans women, and particularly trans women of color, for purportedly loitering for the purpose of prostitution. The horror of being trans while incarcerated, even Chelsea Manning’s gender identity was held as a point of threat, attempting to intimidate her into complying with the state. If you haven’t read it, I highly recommend Going Stealth by Toby Beauchamp.

So in terms of this workshop, sex work stigma has been used to justify research for, and the deployment of surveillance technology that ultimately profiles and further harms at risk communities. Non-consensual image and data scraping of online sex work platforms in order to train AI victim tracing technology, for instance, Stanford bragged about scraping more than 30 million advertisements of sex workers.

There’s zero accountability measures for how these images are classified, are they classified as criminal or victim. There’s no accountability about how they’re actually used, though we get reports of sex workers being stopped at borders or being identified in other manners despite any sort of attempts to actually create and protect their identities.

Algorithmically informed account shutdowns that strip laborers of their sources of income. We don’t know how algorithms codify sex workers. Literally, what does a sex worker look like to a machine or who has defined what high-risk behavior is and how that high-risk behavior can stop that person from being able to access income. The sex worker is both indispensable yet disposable for those behind the development of these technologies

So with all of that heavy information in mind, let me introduce our two panelists.

Kendra Albert is a clinical instructor at the Cyberlaw Clinic where they teach students to practice technology law. They hold a degree from Harvard Law School and serve on the board of ACLU of Massachusetts. They enjoy redistributing money from institutions, working on their solidarity practice and making people in power comfortable. Thank you, Kendra.

Amanda Levendowski is an associate professor at Georgetown Law, where her research and scholarship examine how intellectual property law can be used to creatively address challenging social issues that cross cut privacy and technology. Her recent work has explored non-consensual pornography, biased artificial intelligence, secret surveillance technology, and invasive face recognition.

Thank you both for being here and I will now pass the torch.

Kendra Albert: Awesome. Thank you so much, Gaby. That was an amazing introduction. And I feel really grateful to be here with Amanda and with all of y’all. I am on the unceded lands of the Lenni-Lenape people in Philadelphia, otherwise known as Philadelphia, Pennsylvania.

Amanda and I figured we’d start out by talking a teeny bit about sort of what is trademark and how does it fit into a broader constellation of tech tools, legal and otherwise, that you might use to learn about surveillance that affects your community or that’s happening in your community.

I’m going to turn it over to Amanda, to talk very briefly about what trademark is, because she has that down way better than I do. And then I’m going to talk about a little bit about some of the other tools that you might use. We’ll talk a little bit about why you might use trademark transparency in particular. And then Amanda will start with her very excellent slides.

So Amanda, over to you.

What is trademark?

Amanda Levendowski: So other than that so trademarks are at their core a form of intellectual property that protects a brand. You can think of Coca-Cola as a trademark, you can think of Georgetown University as a trademark, you can think of Harvard University as a trademark.

What trademark does in the context of what we’re going to be talking to you about today is using a tool called the Trademark Electronic Search System, or a TESS just to keep it catchy. And one of the things you can do is search to basically figure out how brands described themselves, to figure out what they’re actually saying when they create a surveillance technology that has a branded image, what do they use to describe its capabilities, its scope, and in some cases, how do they actually market it and advertise it to law enforcement to figure out how it’s going to be deployed in practice?

So that’s what we’re going to be talking about today. This is not how trademarks are traditionally used. This is one of those creative ways to use IP to break the system that Gabriella mentioned in our introduction. But it is a way to creatively uncover how these sort of aliens companies are describing their products in their own way.

How does trademark fit into the broader constellation of tech tools?

Kendra Albert: Thank you so much, Amanda. So I think one thing I want to just start with is to talk about how this fits into kind of our broader constellation of tools that you might use to get information on surveillance.

And I want to start by acknowledging that often the first and best tool to get information on surveillance is talking to the folks who are surveilled, right? About the things that they experience, about the ways in which they have seen things that police or law enforcement or child protective services, or any other sort of institutional actor have sort of found out about them. Because often that’s actually the first invest clue as to what’s going on sort of more broadly behind the scenes.

But let’s say you’ve already done that, what other tools are at your disposal?

So one thing that Amanda will talk about a little bit is procurement processes. So sometimes governments when they’re buying surveillance tools, either off the shelf or sort of asking companies to create them for them, will develop records of either what they’re looking for or what’s actually developed. And you can get those through the sort of lens of looking for procurement stuff. The legal tool you usually use to get that is freedom of information request.

So in the US there’s a Federal Law, which is FOI, the freedom of information act, but each state has its own FOI laws. That can be really useful, especially if what you’re looking for is certain kinds of government information. You can get all kinds of stuff that way.

But FOI laws generally don’t apply to private corporations. So you’re not necessarily going to be able to get ahold of Amazon’s internal information about how ring works using FOI, unless they’ve sort of sold it or given it to a government official and you’re able to kind of navigate the thicket of exemptions, and exceptions, and government delay tactics to sort of get a hold of that.

Other tools that you can use to kind of get access to government information about surveillance include litigation. So if you do sue a government agency over any particular set of practices, when you’re in what’s called discovery, which is sort of a process by which you can request information, you can sometimes get information either through talking to officials on the record, or being able to get documents about sort of different types of surveillance practices. What other methods am I forgetting Amanda?

Amanda Levendowski: There’s two that we’re going to talk about in just a minute, and we’re going to talk about securities and exchange commission filings which is a kind of surprising place. And I’ll explain a little bit more about that in a moment, but just know that they exist and you can find out information about a company’s market practices from those filings.

And the other one is patents, which we will also talk about in more depth than just a moment. But patents are publicly filed. They exist in a publicly searchable database, and there are ways to find otherwise previously unannounced surveillance technologies using patent system practices.

What’s the difference between a patent and a trademark?

Kendra Albert: So maybe you’re going to get to this in the slide. So if so, tell me to hold it and I will. What’s the difference between something in a patent and with a trademark?

Amanda Levendowski: Oh, we will get to it, but I would love to give a little amuse-bouche of what it is. The big difference is trademarks are related to something that is currently being used in commerce, the brand is currently being used on a product or a service, or there’s an intent to use it. You have to have a bonafide intent to use the brand in connection with the thing that you’re saying you’re going to use it for in order to acquire a Trademark Lawfully.

Patents are not a promise to create a product. You could file a patent for something that is technically impossible to create right now, but may exist in the future. You could file a patent defensively to prevent other people from trying to move in on something that you’re not interested in doing, but you also don’t want anyone else doing. You could also file it proactively about something that you may want to develop a long time in the future, depending on a bunch of different factors.

So those are the main differences, is a patent is not a promise to produce a product just to keep it alliterative and a trademark requires you to either be producing or have the intent to produce a product or service with that brand on it.

Kendra Albert: Awesome. So with that, do you want to get into the slides?

Using the Trademark Electronic Search System to Find Surveillance Technologies

Amanda Levendowski: Let’s do it. Okay. So the title of his talk is Fighting Surveillance Tech with Trademark Transparency. I am so grateful to be here. Thank you for that wonderful introduction, Gabriella and Kendra, and it is just an absolute pleasure and I’m so excited to talk about this project with all of you.

And hopefully my goal for the project is that you leave with two things. One, I want you to leave knowing something you didn’t know coming in. And two, I want you to leave feeling confident that you can use the Trademark Electronic Search System to find out something about a surveillance technology. So maybe that’s a really ambitious second goal, but I feel confident that we can do it.

But before we get there, I want to talk about how I came to this project. I came to this project because of Amazon Ring. A couple of almost two years ago now, I went on the Trademark Electronic Search System, I came up in my practice as a Trademark Lawyer when I was working at a private company. And I was curious about sort of the intersection of trademarks and surveillance, because I was like, “Well, you should be able to find out about surveillance technology at…”

Waffles is here… this is Waffles, she’ll be making a guest appearance….

“You should be able to find out about surveillance technologies based on the way companies describe their products.” And so I popped on a test and I looked up surveillance. And it just so happened that Amazon had filed a trademark for a product called Amazon Ring.

This was almost a year before it was publicly exposed that it was being partnered with law enforcement to essentially create a giant surveillance, a civilian surveillance dragnet of these devices, including ones that were sold or not sold, but given to law enforcement to give out to their communities to create a bigger network.

So, in other words, if we had been watching the trademark register more carefully, if we have been using tests more strategically, we could have found out about Amazon Ring a year before it was sort of on the public consciousness radar, and we may have been able to mount more of a defense and a resistance.

So that’s one opportunity that we have by using the Trademark Electronic Search System is finding out about these surveillance technologies before they are already entrenched in our communities.

Using Securities and Exchange Commission Filings to Find Surveillance Technologies

So I mentioned two other places that you can find publicly available information with pretty low friction. You can look through securities and exchange commission filings to find out publicly available information about how Amazon describes its strategic business plans.

However, as you can see, this is pulled from the 2018 acquisition activity filing from their K-filings, and you can see that it doesn’t reveal a lot of specific information about the plans for Ring. It doesn’t tell you that they were going to create, potentially partner with law enforcement with this technology. It doesn’t reveal that it’s going to be used for surveillance in that particular way.

It just talks about how much it costs and why they think that they want to do it, which is to serve customers more effectively. So you can use K filings, but it’s not going to be the most revealing.

Using Patents to Find Surveillance Technologies

Another possibility is patents. This is a form real patent that Amazon filed for generating composite facial images using audio visual recording and communication devices. Essentially this would be technology for again, surveillance.

However, when this was written by the press Amazon’s comment was, “We aren’t producing this product. This is just a patent file, a lot of patents, it’s kind of what we do. So you can’t really take this at face value and say, we’re going to produce a product, ’cause that’s not really how the system works and that’s not how it really how our business operates.”

And while that does sound a little bit like corporate doublespeak, it’s also true. It’s also how it operates in practice. There are a lot of technology companies who have some very sketchy patents, but a lot of those products are never, ever going to come to pass.

That does not mean we should not scrutinize these patents and take them to task for filing in the first place, but it’s not going to be the most revealing source of information because it’s always possible that it is just a perspective patent, not actually a promise to produce the product.

Why is surveillance transparency hard?

So the overview of what we’re going to talk about today is why is surveillance transparency so hard, ’cause it’s pretty hard as a lot of you know from experience, and how Trademark Law can help?

This is definitely not what was intended by the folks who pass the Lanham Act in 1946, but we’re going to take it and give them a new legacy, which is surveillance transparency. So let’s start with why is surveillance transparency is so hard?

I’m about to take you through a journey with the Queens of Obfuscation, who are a bunch of bad-ass scholars who have all the written about reasons surveillance transparency is so darn hard. So we’re going to go with the Queens.

The first queen is Catherine Crump, who’s written about, as Kendra mentioned, problematic procurement practices. I’m just, alliterating all over the place today. One of the challenges is that not every jurisdiction has procurement policies in place.

Procurement policies have basically guidelines for how municipalities and localities can acquire surveillance technology. And in some cases they have requirements that there’d be a public hearing, where civilians could hear about the technology and they can weigh in. There may be requirements about there being a civil liberties oversight board, who brings an opinion about the impact of the technology.

It can take a lot of different forms, but as maybe some of you know, because you’ve never heard of this before, not every jurisdiction has them and not every jurisdiction has them with the same strength.

So that’s one challenge, is while procurement policies can be really valuable in terms of finding out about these surveillance technologies before they’re entrenched, they’re not available everywhere and they’re not available equally.

Our second queen is Elizabeth Joh, who talks about private power. Because as I’m sure a lot of you know, a lot of these law enforcement agencies are not using law enforcement developed surveillance technology in order to use to surveil marginalized communities.

They’re using privately developed technologies by companies like Amazon, like Amazon Ring, in order to exercise their surveillance practices. And because of their private power, that means that a lot of the ways to work transparency that are usually available for government developed products are not available. And that’s one of the things that Elizabeth talks about in her [inaudible 00:18:04] scholarship.

The reason this comes up is because this can get in the way of freedom of information act requests, because our third queen, Hannah Bloch-Wehba, has talked about the flaws with the Freedom of Information Act and local laws in order to find out about these surveillance technologies, because there are lots of ways to obfuscate, and delay, and otherwise be non-responsive, or as Kendra alluded to, exemptions, exceptions, and delays.

What is trade secrecy?

There are also another challenge when it comes to surveillance technologies in particular, and that’s toxic trade secrecy. Trade secrecy is actually the fourth area of intellectual property law, it doesn’t get talked about in the same breath as the others, but it is just as important in the context of surveillance technology and particularly other algorithms because trade secret law is often used to shield those from public scrutiny as Sonia Katyal, our next queen has talked about.

In this context, companies can insulate themselves from disclosure because when you file a freedom of information request, you may not be required to disclose that information as a government agency if the information is protected as a trade secret.

This applies more to algorithms than it does to physical products, but a lot of you mentioned that you’re interested in algorithms generally, and this is one way that you would not be able to find out that information is [inaudible 00:19:22] the Freedom of Information act.

Rebecca Wexler has also written about this in the criminal context, because if there’s contact with the criminal system and an algorithm was involved, trade secrecy can be used to shield that even from a criminal defendant, which is particularly alarming. So all of these queens have addressed the ways in which surveillance transparency is so hard.

How can trademark law help with surveillance transparency?

But luckily I’ll be the next queen, I’ll tell you, there are ways that’s Trademark Law can help. So this is a quote from a case about Trademark Law, and it is unlikely that more than a tiny fraction of the public has any idea what federal registration of a trademark means.

Well, you are now in that tiny fraction of the public, who knows what trademark registration means, or you will by the end of this conversation.

So there are three key disclosures and trademark applications. To back up a little bit in order to acquire a trademark, this is a federal process, you have to fill out a trademark application, submit it to a government agency called the US Patent and Trademark Office, where a trademark examiner will do exactly what the title says, they’re going to examine it. And they’re going to see whether it hits all of the qualifications in order to become an actual federally registered trademark.

Most of the brands that I mentioned earlier are federally registered trademarks. Most of the brands you’re familiar with are also federally registered trademarks. As part of that process in the application, there are key disclosures which are relevant for finding out about a surveillance technology. There are others, but these are the three big ones. And I’ll be honest, they’re the three sexiest ones.

So the first one is a use definition. It doesn’t sound sexy, but I promise it is I’ll deliver. They use designation is what I was saying earlier, that’s the opposite of patents.

In order to get a federal trademark registered, there has to be… The mark either has to be in use, or there has to be bonafide intent to use. In other words, a company couldn’t come back and say after filing a trademark for a dystopian surveillance technology, “Oh, well, we didn’t really mean to do anything with it. We didn’t have any intention of producing the product.”

Because then while you’re in a rock and a hard place, you either lied to the US PTL and made a materialist representation in your trademark application, or you’re lying to the public in a pretty transparently obvious way. You can take your pick, but not either of them are equally… They’re both bad.

The other possibility is in your goods and services description. And this is what I was alluding to earlier, where how the brands have to describe what the thing that they’re on, actually writing it out. So we will look at some examples later, but just keep in mind that a goods and services description is basically a TLDR of what the thing does. It’s just written out in a fancy way. That’s all it is.

And the third possibility of finding important information about surveillance technology in a trademark application is held in what’s called a specimen. After you file your trademark application, if it’s in use, you have to include a specimen that shows how the mark is being used.

And that includes photographs of a product, that includes the interface of a computer software system, that includes marketing materials, in some cases, that includes a really sensitive information, developer guides as we’ll learn later. And those are all ways to show how the mark is used in commerce. And this is probably the most fruitful way to find out about surveillance technology, because it is often the most revealing and it’s the most straightforward.

What does a trademark application look like?

Now I just gave you a lot of information. We’re going to go through it and it’s going to be easier to understand and practice. But I want to come back briefly to Amazon Ring and show you what it looks like to find out what that goods and services description looked like that revealed, to me at least, that I thought it was going to be used for law enforcement down the line.

So this is a screenshot of the Trademark Electronic Search System, our buddy TESS. And I just searched for the term surveillance in the field goods and services.

This is what a trademark application looks like. I’m going to move. Can you see my mouse on the screen? Great.

So this is in class nine, class nine, which is for computer hardware and software. So there’s international classes that essentially categorize the different ways in which the products are explained. And one of those is saying that this is hardware software, which makes perfect sense. It’s a little bit of both. But I highlighted the part that stood out to me, saying that this is an electronic surveillance device that can be deployed to gather evidence or intelligence.

When you’re talking about a smart commercial doorbell, that’s just generally not how you describe your product in my opinion. I’m not an expert. I’ve never written the goods and services for a commercial doorbell, but I don’t think that if it was just going to be a commercial doorbell, I would describe it with the language evidence or intelligence.

To me as a lawyer those are lawyer words that signaled lawyer things and they specifically signal a particular type of law enforcement surveillance. Because who are the kinds of people that gather evidence? Law enforcement. I mean, other people too, but we’re not all in the law homes.

So in the paper that I wrote, that I accompanied this theory of how we can use trademarks for surveillance transparency, I explored three pretty significant surveillance technologies.

The first is PredPol, which is a predictive policing algorithm. The second is Stingray devices, which are used to intercept cell service, and they’re often used at protests. And the third is Vigilant Solutions, which was used for automated license plate readers.

However, I’m not going to bore you with all three of them. If you want to go read an academic paper about this theory, please be my guest, but we will be adapting it into a more TLDR friendly format as well.

What is PredPol?

I’m going to start with PredPol though, because this is probably the best example from the paper that shows the power of the Trademark Electronic Search System for checks or surveillance transparency.

So dozens of cities had secretly experimented with predictive policing software and public record requests were used to verify some of those departments. So this is a case of working in tandem with freedom of information laws in order to amplify the impact of the surveillance revelations.

So if we go back to TESS, this is the interface for PredPol. You see class nine again, which makes perfect sense. Class nine is very commonly used for surveillance technologies because again, that Is hardware and software.

Class 42 is another class that is commonly used, it’s related to computer and scientific instruments or services. And so you will see those two coming up again and again in your searches. And you may even want to explore within those classes for particular terms to keep a really broad approach to your investigation of these surveillance technologies.

But this doesn’t really tell us anything we didn’t already know, because we did know that PredPol existed. This isn’t like Amazon Ring, where we could have had a headstart on this technology.

So what is so revealing about the PredPol application? Well, I told you that there was a specimen file, that there has to be a specimen filed in connection with the trademark application, either when the mark goes into use, and you can see that there was a statement of use on the same day and they submitted something kind of interesting for their specimen.

This is the third page of their specimen. They decided to share an actual contract with a jurisdiction that had not been publicly disclosed yet. And we know that because that hadn’t been one of the jurisdictions that came up in other journalists searches for municipalities that were using PredPol. That was Red Richmond, California.

And there’s some pretty interesting information in here, how much money it costs and the fact that they’re getting a discount, or in part G that there will be engaged in joint marketing, including but not limited to press conferences and media relations, training materials, marketing, trade shows, et cetera.

And my favorite part, for example, if a chief is required to attend and speak at a conference of police chiefs, to which they are not already traveling, PredPol agrees to reimburse the city for travel expenses if requested.

They also included somebody’s real cell phone number, which I redacted because I’m sure this guy doesn’t know that this is in the trademark filings, but I’m sure he would not be thrilled to find out that his employer included it.

But this is the kind of information that these companies accidentally reveal. I’m not a sports person, let’s be honest this is absolutely an own goal. This did not have to happen. It could have uploaded a ton of different stuff, but they happened to upload this contracts. And it’s particularly rich because well, those joint marketing efforts didn’t work out, because the Richmond police chief said that they were going to discontinue predictive policing software use midway through their term.

So we’ll do I think, are we going to do questions and comments Kendra? Or do we want to…

So we’ll do questions and comments, and I hope that that was enough to give you a little taste of what we’re going to explore today before we turn it over for… I’ll do one live demo, and then we’ll let you go wild and figure out what you can find.

Kendra Albert: Awesome. People should feel free to like unmute and ask questions if they’re comfortable being on the recording, or feel free to either throw them in chat or just message me directly and I will happily read them. But questions, comments, reactions before we sort of dive in.

Amanda Levendowski: We can also just dive. Do you want to dive? Thumbs up if we want to dive? Okay. So let me stop sharing and I’ll show you a live demo to give you a sense of what this will look like.

Kendra Albert: And I think we can keep recording through the live demo and then stop after. Does that sound okay to you, Amanda?

Amanda Levendowski: That sounds good to me.

Kendra Albert: Perfect.

What does the Trademark Electronic Search System (TESS) look like?

Amanda Levendowski: This is what a standard search on the Trademark Electronic Search System, our buddy TESS looks like. We will drop a link in the chat to the US PTO website. So you can go there and go straight to TESS yourself and have it open on your computer. But here is the place to start.

So this is a regular search, this is your basic search. And you can put in the term Amazon Recognition, which is Amazon’s commercial enterprise facial recognition software, which is used by a number of organizations in ways that could identify as sex workers.

So you clicked submit query, just kidding.

Kendra Albert: The problem with live demos always.

Amanda Levendowski: I just did that to make everyone feel great if they had troubles later. This was just normalizing those mistakes and just keeping it iterative. So you submit query.

So this is what an actual page looks like that’s not cropped or a filing like this. It has goods and services descriptions of what Amazon Recognition is, image analysis and identification, animal recognition. We know that that’s not really the concern that we have, it’s the face recognition in particular.

But this gives you a sense in class nine, as we previously discussed, class 42, and it has the holy triumvirate of classes likely to disclose surveillance technology, class 35. And those are again, just ways to chunk out how these brands are describing themselves.

What kinds of categories of goods and services they are associating themselves with. And so they’re telling us through classes nine, 35 and 42, that is as hardware and software, computer and scientific services and entertainment business services, that’s class 35 that are going to be relevant.

So you can continue that information as you go through your searches. And we will remind you of that as well. So now we’re going to leave TESS and go to another service, but you don’t need to know the name of it. It’s called TSDR, that’s the TLDR for the TSDR.

We’re going to wait for that to load.

So now we’re in TSDR and you can look at a lot of that information in just a different way. So what you’ll do is you’ll navigate to the documents tab and you’ll navigate to the specimen, because that’s going to be where you’d find the beefy information. And one of the things Amazon uploaded is the developer guide for Amazon Recognition and they uploaded the entire thing.

So it is completely possible that they could find… Gabrielle was like, “Oh shit.” And it’s terribly true. This is a huge question about whether or not this is something they meant to disclose.

I’m sure it does adequately show the Amazon Recognition mark in association with the relevant goods and services, but did they need to disclose the developer guide in order to get that? No, they did not.

This is another own goal, this didn’t have to happen, but it’s for our benefit that it did. However, now that we’ve done the live demo, you can see the basics of navigating the Trademark Electronic Search System.

This is the kind of thing that you can all go and discover for yourselves. And so maybe we can offer questions again, and then if there are no questions, we will go into breakouts and people can explore the system.

Where do you get started with trademark transparency?

Kendra Albert: So there are two great questions from chat. One is, “What do you do if you don’t even know the name of a thing or what to start even looking for?” So I’ll start with that one.

Amanda Levendowski: That is such a good question. Where to begin when we don’t know what the brand name is. We begin with goods and services, because as you recall, I did not find out about Amazon Ring because I searched for creepy Amazon stuff. Although, could I? Yes.

I did it because I searched for just surveillance in the goods and services. And there are lots of goods and services that may be relevant. I did do some work earlier to see what kinds of goods and services were relevant potentially specifically related to sex workers.

You will be interested to find, I don’t know, maybe you won’t be, but I was interested to find that there happened to be only three trademark applications or registration that used the phrase sex workers.

One is for a denim company, and I have some questions about that. And then one is for a charity and then one of them I already forgot. But they’re not anything related to surveillance.

There is a specific place where you can look for particular phrasings of goods and services descriptions. Are they official phrasings of the US PTO? Basically it’s called the trademark manual of examining procedure. That’s one place where you may find examples.

There’s also the trademark ID manual, which I think I have up, yes. The trademark ID manual, you can also look and see, this is in the Google doc. You can look and see for particular phrases or particular words that might be relevant that have been pre-approved by the US PTO and they are likely to be more commonly used as goods and services descriptions because they’re pre-approved, and that means it’s easier for an examiner to look at them and say, “This makes sense. I’m good to go.”

So you can experiment with this tool as well to come up with goods and services descriptions related to surveillance of a whole variety of communities, individuals, but you may also want to focus on the technology because some of the most dangerous technologies for marginalized communities are face recognition technologies. Those can be particularly invasive, particularly damaging and particularly dangerous.

So looking at face recognition technologies by commercial companies may be another way to find out about the technology without knowing its name. So goods and services is the ticket when we’re talking about finding out about things when we don’t know the name.

Kendra Albert: Okay. Amanda, could I ask you to do me a favor and just do a very brief, the same thing demo, where you searched for goods and services.

Amanda Levendowski: I guess.

Kendra Albert: We can even give you a phrase if you want.

Amanda Levendowski: Oh, I was going to do the only one that came up earlier related to… Well, give me a phrase.

Kendra Albert: Fraud prevention. Ooh.

Amanda Levendowski: Interesting.

Kendra Albert: Fraud prevention, yes.

Amanda Levendowski: So the ones that are crossed out means they’re no longer valid, but that doesn’t mean that the goods and services description they were used for initially that product doesn’t still exist. But forensic analysis of surveillance video for fraud and threat prevention purposes, that’s one. Check marking machines on preventing fraud, and let me just even take this to another level.

We can go back to TESS, and we can go to a structured search. Now we’re elevating, we’re taking it to the next level. Okay, we’re ready for it. Fraud and threat prevention purposes, that’s a phrase from the verbatim one, you change it to goods and services, and then you submit the query.

These are all the technologies that have been registered at the US PTO that use that phrasing in their product. So you may have some false positives that happens occasionally, I did put it in quotes so it should be exact. What you can do is start here and go through these and see if any of them are relevant.

So for example, Stealth Monitoring for some reason it seems like it might be promising. Here you go. They do the forensic analysis of surveillance video for fraud and theft prevention purposes.

So now we know this is a company, Stealth Monitoring, it’s based in Texas, and we actually know a good amount about it. We know a lot about what they say, what they do. Let’s go to TSDR.

I know we’re all waiting with bated breath.

Kendra Albert: I am, I stopped breathing a while ago. I’m very excited.

Amanda Levendowski: What is the specimen going to be? Is it going to be weird? Who knows? Okay. So this is actually a pretty standard specimen, although it could potentially reveal sensitive information.

I haven’t looked through all five pages, as you all know, this isn’t a magic trick. I didn’t like know in advance that Kendra was going to give this one to me.

But this looks like a pretty standard printout of marketing materials that show the Stealth Monitoring mark used in commerce with those goods and services related to forensic analysis of footage from video surveillance to look for fraud and theft.

So that’s how you could reverse engineer finding out about a company or finding out about a product when you’re just operating from using the goods and services.

Kendra Albert: Great. So there are two more questions.

Amanda Levendowski: Great.

Can you search by trademark class or groups of classes?

Kendra Albert: “Can you search by class or groups of classes?”

Amanda Levendowski: You cannot search by groups of classes because the US PTO is pretty old school, but you can search by class and you can also search by goods and services within the class.

As I showed you with the structured search, that next level searching mechanism, you can search by one line being the class, just remember that you have to put two zeros before a one digit number, I don’t know why, but otherwise it won’t work. Two zeros and then class nine and you can search for it.

And then the other thing is you can also search for the goods and services descriptions in that class. So you can search with an and joinder and you will have a more specific search than if you’re just searching in one.

So you could literally just go through and search for all of the surveillance marks in class nine and 42, if that’s what you wanted to do, however, they would have to be two separate searches because you can’t search multiple classes.

What would stop someone from creating a trademark scraper for a particular good and service?

Kendra Albert: And then one last fantastic question, “What would stop someone from creating a scraper that grabbed all of the marks with a particular good and service?”

Amanda Levendowski: Nothing. And indeed, I’m delighted to tell you, thanks for teeing that up, a student at Georgetown and I, Dylan Brown Bramble have actually developed a tool called Trademark Watcher that does exactly that.

It essentially monitors for surveillance related marks that show up on the US PTO’s registered and we’ll send eventually either a tweet or an email digest, we’re still experimenting with what we’re going to do with it. And it will basically give you an update about here’s what’s new in surveillance tech.

So the US PTO’s website is very scrapable is it turns out and we’ve tried to do it so that it would go off late at night and it wouldn’t interrupt any services because there’s another law that we’re not getting into that could potentially be an issue.

So the goal is to basically make this process easier. I think it’s great for everyone who’s in this workshop to learn how to do this for yourself, so that you can experiment and find things that the bot may miss.

But if you’re looking for a new technique to have in your repertoire to keep an eye on surveillance technologies coming down the pipe, this could be a really powerful tool. And also anyone could create a competing or other tool, a complimentary tool.

It’s hard. I didn’t do it Dylan Brown Bramble coded it, but it’s doable. So you could also do this for yourself if you were curious in particular technologies.

Kendra Albert: I swear I actually did not tee that up in advance. That was an actual question from the actual chat. So before we stop recording and we send everyone into sort of doing this for yourself, any last questions? We’ll obviously provide materials related to all of this. So there’s plenty more coming. But any last questions before we stop?

Amanda Levendowski: Okay.

Kendra Albert: Awesome. All right.

Leave a Reply