#9 Facial Recognition In Focus

Bad Banjo // Tech Jukes on Facial Recognition // TikTok Sucks

Surveillance takes many forms. You may immediately think about things like facial recognition, browser history tracking, or perhaps location data tracking as the most obvious and scary examples. Indeed, facial recognition software has grown as such a concern in the public consciousness that John Oliver was able to run a segment on the topic:

Surveillance goes so much deeper than cameras and location, however, and I want to touch on Banjo, the data analysis company that recently had to shut down after its CEO, Damien Patton, was exposed as a Nazi.

Banjo’s main product was in real-time analysis of city data to predict when and where certain events would occur. Any pre-cog crime prediction tool like this obviously requires a significant data set, and it appears that state government officials in Utah had no problem handing over huge quantities of data about American citizens to a private company.

The agreement gave Banjo real-time access to about 1,000 Utah Department of Transportation surveillance cameras, 911 emergency calls, and some location information for first-responder vehicles.

Patton isn’t just a standard-issue anti-semite, he was actually involved in the a drive-by shooting of a Synagogue. Banjo had raised $220M from investors including SoftBank and had won several contracts with US states when Matt Stroud and David Gershgorn broke the story in Medium’s OneZero publication. Stroud’s piece on April 28th caused immediate blowback against the company (OneZero). The next day, Banjo began losing contracts (OneZero) and soon thereafter, the company suspended operations (OneZero) and Patton was forced to resign (OneZero).

News like this touches on another hot button issue of the day, namely law enforcement and how they receive and spend their budgets. The Overton Window has shifted so far in the United States that datasets like these sound tame to us, but private corporations getting access to civic data is something worth talking about. It’s never too late for us to stop, ask questions, and reconsider whether giving data like this to private corporations is actually in the best interest of citizens, or a valid use of tax funds.

The protests that have consumed the nation in the wake of George Floyd’s death have been a turning point in our civic discourse unlike anything I’ve seen in my lifetime, but police forces are already using facial recognition technology to identify and target citizens at these protests (OneZero). This technology is flawed, however, and we’re already starting to see false arrests as a consequence of leaning too heavily on computation to fight crime (NPR). As a nation, we don’t even have a good understanding of how often these systems fail and get it wrong (OneZero).

The potential harm that these systems can cause are leading to calls to completely ban facial recognition from modern cities (OneZero). It’s also causing scrutiny from Congress (House.gov). Boston has now moved to ban facial recognition technology city-wide (WBUR, BuzzFeed). Given the amount of uncertainty, lack of trust, and proven track record of abuse around this technology - I believe slow caution is the right approach. The world has not grown any more fundamentally unsafe since 9/11 as far as I can tell, and yet the “crime fighting” technology we employ has advanced by leaps and bounds, resulting in mass surveillance with little benefit to society.

Some tech companies are starting to wake up to the realities around law enforcement’s abuses of facial recognition technology, or at least the risk it poses to their bottom line given public perception. Amazon is placing a 1-year moratorium on selling its facial recognition tech to police departments (Reuters, Amazon, The Verge). This wasn’t all from the goodness of their heart however, and is also the result of a multi-year campaign by privacy advocates to get the company to reconsider its position (Technology Review). IBM appears to be moving away from facial recognition tech entirely (The Verge), but given that the tech is now off-the-shelf, this might be more of a business decision than an ethical one.

Facial recognition continues unabated in China, and with the blessing of the US government (OneZero). Double standards abound, and they’re worth calling out. For now, we have higher standards of privacy in the US than China enjoys, but it’s clear that the position of those in power is that this is one more item in their toolkit for control, and that they won’t hesitate to deploy it to maintain power. Remember when we thought having your identity stolen was a big issue? Now, your inability to change or lose your identity might be an even bigger risk, if you happen to fall out of favor with your local authoritarian state. You only get one face in life. Who owns access to it?

Bits of News:

  • 🇺🇸 US Senators remain convinced that citizen-level encryption is a bad idea and they seem determined to undo it. If we lose this, we lose free speech online (The Verge).

  • 🦊 Mozilla with a piece praising Apple’s decision to allow users to disable IDFA (tracking across sites/apps). Apple really does seem to be the most privacy-forward of the big tech companies, and I’m glad to see them beefing up their security models even more (Mozilla).

  • 🚫 Bytedance has cut off its domestic Chinese engineering teams from accessing data from international users of TikTok and other digital products. The deep irony here of course is that they had previously alleged that this data was already blocked from their Chinese teams. I guess it’s never too late to modify your truth (PingWest).

  • 🛰 Facebook has taken a 10% stake in the Indian telco Reliance Jio Platforms. This is a move that should terrify anyone concerned with combatting monopoly power as Facebook makes its move from social network to global data network (The Information, TechCrunch).

  • 📱 TikTok has been busted for capturing users’ clipboard text (Forbes). As tensions escalate between India and China, India has banned TikTok entirely (TechCrunch).

Ad Hominem is a blog and newsletter about Ad Tech, Privacy, and Surveillance Capitalism. It currently has no paid subscription. If you want to support this project, the best way is just to spread the word via Twitter or forwarding this email to a friend. You can also send me a tip on my BuyMeACoffee page.

If you have any feedback on this or any other issue of Ad Hominem, please let me know on Twitter. Thank you for your time and attention. I know it’s your most precious resource.

— Sam