AI Initiative
The Ethics + Governance of Artificial Intelligence Initiative



Guest Blogger: Amanda Levendowski - "How can we learn about the AI systems that might be used to surveil us? The federal trademark register has answers"

Our new open call for ideas, AI and the News: An Open Challenge, is looking for submissions that address the impact artificial intelligence has on the news and information ecosystem. Here, researcher Amanda Levendowski proposes one fascinating method for uncovering the hidden details of the AI systems that watch us. - Tim Hwang, Director

We know very little about the artificial intelligence (AI) systems that watch us.

Take Amazon Rekognition. The AI service was announced back in 2016, but there wasn’t a public debate about how its real-time face recognition capabilities could be abused until the ACLU of Northern California, coordinating with affiliates in Florida and Washington, obtained public records revealing that law enforcement agencies were using it in their investigations.

We deserve to know about AI systems that can affect our civil liberties. But we’ve already seen trade secret law, copyright law, and private contracts all invoked to shield AI systems from public scrutiny. And we know that outside funding or in-kind donations - instead of public sources - can make the acquisitions harder to track.  Too often, we don’t find out about AI systems that can be used for surveillance until after those systems are built, acquired and deployed.

It’s impossible for the public to engage in discourse about whether an AI system is fair, accountable, transparent or ethical if we don’t know that an AI system is being used to watch us—or if we don’t know the technology exists at all.

We need a free, public and easily accessible source of information about the AI systems that might be used to watch us. And I’ve found it in a surprising place: the federal trademark register.

The Trademark Electronic Search System, or TESS, is a free and easily accessible database of millions of pending, registered, cancelled, expired or abandoned trademarks. Some journalists already use TESS, which is run by the U.S. Patent and Trademark Office, for their work, but we’ve never had a systematized approach to using trademark filings to further the public good.

I’ve been researching how disclosures in trademarks filings can be used by journalists, researchers, technologists and civil society to enhance existing transparency and accountability mechanisms and fuel new ones for my next article, Trademarks as Surveillance Transparency. Federal trademark filings demand extensive public disclosures, and searching TESS for those filings offers a crucial advantage over other free, public sources, like representations to regulatory bodies or disclosures in patent applications: it requires no legal or technical expertise.

In fact, searching TESS is easy.                                            

If I want to investigate AI systems targeted to certain customers, I can run a search for filings that include the word “AI” and a certain phrase. If I choose “law enforcement,” I’ll find a filing for the AXON AI mark, which covers in part “[m]otion picture films that include evidence of incidents in the fields of law enforcement, military, professional security, first responders, and consumer security.” The application was filed on Feb. 16, 2017 by Taser several months before the Taser rebrand to Axon was announced publicly.

Or I can search by intended purpose. When I searched filings with the word “surveillance” last week, for example, I turned up a new filing from Amazon for AMAZON RING, which covers in part “[a]utomated self contained electronic surveillance devices that can be deployed to gather evidence or intelligence,” along with a long list of dozens of other uses.  It isn’t entirely clear what Amazon envisions for AMAZON RING, but we have time to start finding out. Amazon filed the application just last month, on Sept. 13, 2018.

In the meantime, we can start correcting the longstanding inequalities between the watchers and the watched by using trademark filings to watch back.

Levendowski is a Clinical Teaching Fellow with the Technology Law and Policy Clinic at NYU Law, where she is also an Engelberg Center Fellow and an Affiliate Researcher with the Information Law Institute. In 2019, she will join Georgetown Law to direct and teach a new intellectual property and information law clinic.

Tim Hwang