ClearView AI Will Probably Launch An App, Body Cam = ALPR for Faces, Guidelines for Smart City Surveillance
Welcome back! In this email, I argue that Clearview is going to launch a consumer app, I offer some brief thoughts on bodycam facial recognition tech, link to some documents related to smart city surveillance, and drop a link to info about pending surveillance tech legislation in Maine.
FACIAL RECOGNITION
As Clearview continues to get pounded by facial recognition opponents, their rational move is to launch a consumer app. That’s my thought for the day. Before I get into it, here’s some background from the New York Times:
“Before Clearview Became a Police Tool, It Was a Secret Plaything of the Rich. Investors and clients of the facial recognition start-up freely used the app on dates and at parties — and to spy on the public.” — People (like some people on Twitter, not a representative sample) seem shocked to learn that Clearview put their app in the hands of lots of users other than the police and let those users run around testing it in public. This is only surprising if you’ve never developed a commercial product — of course Clearview tested it in the real world, they needed to make sure it worked before they tried selling it to customers. And it looks like it worked pretty well.
One Tuesday night in October 2018, John Catsimatidis, the billionaire owner of the Gristedes grocery store chain, was having dinner at Cipriani, an upscale Italian restaurant in Manhattan’s SoHo neighborhood, when his daughter, Andrea, walked in. She was on a date with a man Mr. Catsimatidis didn’t recognize. After the couple sat down at another table, Mr. Catsimatidis asked a waiter to go over and take a photo.
Mr. Catsimatidis then uploaded the picture to a facial recognition app, Clearview AI, on his phone. The start-up behind the app has a database of billions of photos, scraped from sites such as Facebook, Twitter and LinkedIn. Within seconds, Mr. Catsimatidis was viewing a collection of photos of the mystery man, along with the web addresses where they appeared: His daughter’s date was a venture capitalist from San Francisco.
“I wanted to make sure he wasn’t a charlatan,” said Mr. Catsimatidis, who then texted the man’s bio to his daughter.
Ms. Catsimatidis said she and her date had no idea how her father had identified him so quickly. “I expect my dad to be able to do crazy things. He’s very technologically savvy,” Ms. Catsimatidis said. “My date was very surprised.”
Clearview’s initial beta testers (some of whom were people using the app at parties and in bars) sure seemed to like the software. Today, Clearview claims to focus on police departments and companies/organizations with security needs. That’s a standard startup business decision — pick a narrow market you understand (and their advisors understood the security/law enforcement market), work out kinks, get user feedback, establish product-market fit. But the next standard move after product-market fit is to expand market share. There’s no rational economic reason for Clearview to limit their business to that initial niche market (assuming the product works and is ready for consumer adoption). In fact, if Clearview is acting in their own interest they should expand their market and quickly. The consumer TAM is much bigger, the barriers to entry are lower, and the landscape around them is quickly shifting.
About that shifting landscape. The niche market of law enforcement and security companies is getting pounded by advocates who are opposed to the proliferation of facial recognition technology. As controversy grows, public sector adoption is being chilled, cutting into Clearview’s bottom line (city councils are imposing moratoriums, AGs are beginning investigations, etc.). The rational move for the company is to expand its target market from the niche law enforcement/security market (with all of those “oversight hassles”) to the much more difficult to regulate market known as “everybody.” Pop that genie out of the bottle and it will be nearly impossible to put back in. 🧞♂️ Bet on a publicly available Clearview app launch.
🏛 While we’re on Clearview, Senator Markey sent a letter to the company questioning whether they are selling their technology to authoritarian regimes. To state the obvious, that’s a bad business to be in. It’s much better to be in the business of merely “indexing images” ala Google and selling an app to consumers ala Instagram. Maybe Clearview ends up doing both, but on the optics, it seems like pivot time, and I bet Hoan Ton-That’s got an app for that.
Live Facial Recognition Is Coming to U.S. Police Body Cameras- Dave Gershgorn has an exclusive article discussing Wolfcom, a company that is pitching body cameras with live facial recognition for law enforcement.
My immediate thought was “This is like an ALPR for faces.” It’s going to raise similar issues with false positives, there will be massive geolocated databases of information (of course the information is more sensitive, faces, rather than plates), data access and retention issues are at stake, and there will be obvious calls for “policies and procedures.” Those are ALPR type things, and like ALPRs, this tech will have benefits that will make it hard for opponents to stop. After all, why wouldn’t a community want their police officers to know that a person they are looking at has an outstanding warrant for a violent sexual offense? Or that a person is a missing person, possibly in mental distress?
Yes, the system, like all systems may be prone to errors (but probable cause assumes errors). Moreover, the argument will be that this is a tool with life-saving potential, helping officers make split-second decisions about whether someone is a threat, helping them make decisions about whether or not to arrest someone, whether or not to provide them aid, etc. One side of the debate will argue for a ban, the other side will argue for reaping the “life-saving benefits” while regulating with “policies and procedures.”
If you want to know who wins that debate, look to how the ALPR debates have played out. 🤷♂️
On the topic of policies and procedures as an alternative to an outright ban, check out what Evan Selinger has to say:
So "instead of describing new regulation as an impediment to business and government interests, they appeal to it hoping for two things," Evan Selinger, a philosophy professor at Rochester Institute of Technology, said in an email.
First, police pushing for regulation "gain goodwill for seeming to be in favor of emotionally palatable outcomes," like responsible use. Second, they "steer the regulatory conversation in the direction of conservative change," he said. That could ultimately lead to policies that don't create meaningfully restrictive rules for law enforcement.
Some, including Selinger, warn that broader deployment of facial recognition technology by law enforcement could lead to a surveillance state — so its use by police and government should be banned. However, regulatory proposals to date largely focus on setting up a framework for responsible use.
He’s right, that’s the playbook, and it’s a winning one.
SURVEILLANCE CITIES
📕As cities adopt surveillance technology like ALPRs, facial recognition analytics, and high-resolution cameras, questions remain about their efficacy. Some scholars from the Urban Institute researched how to increase the efficacy of surveillance systems and published a new report. There’s lots of interesting information in the report about what works and what doesn’t in surveillance tech. Here’s the report: “Optimizing Public Surveillance Systems for Crime Control and Prevention: A Guide for Law Enforcement and Their Municipal Partners”
outlines eight steps for maximizing public surveillance programs’ impact on crime control and prevention. Though it discusses new surveillance technologies available to law enforcement, it focuses on the steps decisionmakers can take and questions they can consider to optimize their surveillance systems. It can also help agencies identify surveillance goals, consider their systems’ limitations and constraints, and develop strategies for meaningful improvements.
📕RELATED: The World Economic Forum just released A Framework for Responsible Limits on Facial Recognition Use Case: Flow Management.
LEGISLATION
Maine Advances Bill to Foster Surveillance Transparency - GovTech covers a new bill moving forward as Maine’s public safety commissioner acknowledged for the first time recently that state police use facial recognition scans as part of some criminal investigations.