Why the FTC is forcing tech firms to kill their algorithms along with ill-gotten data
The Federal Trade Commission is punching right at the heart — and guts — of how data collection drives revenue for tech firms: their algorithms.
“I anticipate pushing for remedies that really get at the heart of the problem and the incentives that companies face that lead them into the illegal conduct,” FTC commissioner Rebecca Slaughter told Digiday in an interview last week.
Slaughter pointed to two cases that reflect what we might see more of from the agency. When the FTC in May settled its case against Everalbum, maker of a now-defunct mobile photo app called Ever that allegedly used facial recognition without getting people’s consent, the agreement featured a new type of requirement that addresses the realities of how today’s technologies are built, how they work and how they make money. Along with requiring the firm to obtain express consent from people before applying facial recognition to their photos and videos and to delete photos and videos from people who had deactivated their accounts, the FTC told Everalbum there was another “novel remedy” it must abide by: it would have to delete the models and algorithms it developed using the photos and videos uploaded by people who used its app.
Put simply, machine-learning algorithms are developed and refined by feeding them large amounts of data they learn and improve from, and the algorithms become the product of that data, their functions being a legacy of the information they consumed. Therefore, in order to make a clean sweep of the data that a company collected illicitly, it would also have to wipe out the algorithms that have ingested that data.
Cambridge Analytica case laid groundwork for algorithmic destruction
The Everalbum case wasn’t the first time the FTC had demanded a company delete its algorithms. In fact, in its final 2019 order against Cambridge Analytica, alleging that the now-infamous political data firm had misrepresented how it would use information it gathered through a Facebook app, the company was required to delete or destroy the data itself as well as “any information or work product, including any algorithms or equations, that originated, in whole or in part, from this Covered Information.”
Requiring Cambridge Analytica to delete its algorithms “was an important part of the outcome for me in that case, and I think it will continue to be important as we look at why are companies collecting data that they shouldn’t be collecting, how can we address those incentives, not just the surface-level practice that’s problematic,” Slaughter told Digiday.
The approach is a sign of what companies in the crosshairs of a potentially more-aggressive FTC could have in store. Slaughter said the requirement for Cambridge Analytica to kill its algorithms “lays the groundwork for similarly employing creative solutions or appropriate solutions rather than cookie-cutter solutions to questions in novel digital markets.”
Correcting the Facebook and Google course
It’s not just Slaughter who sees algorithm destruction as an important penalty for alleged data abuse. In a statement published in January on the Everalbum case, FTC commissioner Rohit Chopra called the demand for Everalbum to delete its facial recognition algorithm and other tech “an important course correction.” While the agency’s previous settlements with Facebook and Google-owned YouTube did not require those firms to destroy algorithms built from illegally-attained data, the remedy applied in the Everalbum case forced the firm to “forfeit the fruits of its deception,” wrote Chopra, to whom the FTC’s new reform-minded chair Lina Khan formerly served as legal advisor.
Slaughter’s stance on forcing companies to kill their algorithms, also addressed in February in public remarks, has caught the attention of lawyers working for tech clients. “Slaughter’s remarks may portend an active FTC that takes an aggressive stance related to technologies using AI and machine learning,” wrote Kate Berry, a member of law firm Davis Wright Tremaine’s technology, communications, privacy, and security group. “We expect the FTC will consider issuing civil investigative demands on these issues in the coming months and years.”
Lawyers from Orrick, Herrington and Sutcliffe backed up Berry’s analysis. In the law firm’s own assessment of Slaughter’s remarks, they said that companies developing artificial intelligence or machine-learning technologies should consider providing people with proper notice regarding how their data is processed. “Algorithmic disgorgement is here to stay in the FTC’s arsenal of enforcement mechanisms,” the lawyers said.
More in Media
Digiday+ Research: Publishers expected Google to keep cookies, but they’re moving on anyway
Publishers saw this change of heart coming. But it’s not changing their own plans to move away from tracking consumers using third-party cookies.
Incoming teen social media ban in Australia puts focus on creator impact and targeting practices
The restriction goes into effect in 2025, but some see it as potentially setting a precedent for similar legislation in other countries.
AI Briefing: Amazon’s new Nova models boost AI model efficiency, accuracy and variety across AWS
One of the most buzzy debuts was Nova, a suite of six new AI models that include understanding and creating text, images and videos.