Meta Settlement Reaches the Features, Not the Engine
CDD Statement on Meta's Settlement with State Attorneys General
FOR IMMEDIATE RELEASE
August 26, 2026 | Center for Digital Democracy, Washington, DC
Contact: Katharina Kopp, kkopp@democraticmedia.org
Jeff Chester, jeff@democraticmedia.org, 202-494-7100
The following statement is attributed to Katharina Kopp, Ph.D., Deputy Director and Director for Policy, Center for Digital Democracy:
“The settlement reaches the features of Meta’s platforms but not the engine that drives them. Meta earns nearly all of its money from advertising. Young people’s attention is the inventory it sells. Design that captures and holds that attention is not a flaw in the business model, it is the business model working as intended. Meta can remove “like” counts. What it will not remove is the reason it built them. The penalty does not change that calculation. Roughly $17 billion paid over ten years is about $1.7 billion a year, or roughly ten days of Meta’s annual profit. Children were harmed for years while this business model ran unchecked. This settlement addresses some of what that produced. It does not touch what produced it.
“The practices at issue have moved on since these cases were filed in 2023. AI is now embedded across the whole of the advertising apparatus, and Meta is among the most aggressive adopters. CDD documented that build-out in our June 2026 report, The Persuasion Pipeline. These systems infer in real time what moves a young person, generate the message to match, deliver it inside a feed, a chatbot conversation, or an assistant’s recommendation where it does not read as advertising at all, and measure the result to sharpen the next attempt. Much of this runs on inference and cohort-level prediction rather than on stored records about an identified child, which is why laws written to govern the data a company holds keep missing it.
“A further weakness is structural. Much of the relief depends on Meta correctly identifying which of its users are under 13 and under 18, a determination made by systems the company builds and operates, with a commercial interest in the answer. Checking that a feature was switched on is one thing. Looking inside the systems that decide who is a child in the first place is another, and that is what is needed here: independent scrutiny, by people outside the company, with real access to how those systems work. More fundamentally, protections for young people should not hinge on what a company claims to know about any individual child. Accountability should turn on what a system does, and where a system’s outputs reach minors in disproportionate numbers, the burden should sit with the company to show it is not running prohibited practices against the children it reaches.
“None of this takes away from the changes Meta has agreed to make, which we welcome. Shutting Instagram down overnight for minors’ accounts, capping daily use, ending notifications during the school day, removing like counts and plastic surgery filters, and new guardrails against predatory adults are protections children should have had years ago. The attorneys general secured them under the threat of a jury verdict, after a decade in which Meta declined to act on its own.
“But protections won this way, one company and one feature at a time, cannot keep pace with what is being built. States and Congress have to reckon with the business model itself and the risks it generates. That means capability-based rules limiting what these systems may do to young people, not only what data companies may hold about them, and an end to advertising and marketing aimed at minors.”
The Persuasion Pipeline: How Commercial AI Targets, Shapes, and Sells to a Generation of California’s Children and Teens is available at https://democraticmedia.org/reports/the-persuasion-pipeline .
###