Policymakers apply social‑media addiction lessons to AI safety for children

The social‑media reckoning sets a precedent

The 2023 Senate Commerce Committee hearing on social‑media harms marked the first time Congress treated platform addiction as a national policy issue. Thirty‑nine experts, parents and former tech insiders testified, and former Facebook employee Frances Haugen warned that “we need to protect children before the damage is done.” The public outcry that followed forced the FTC to issue new privacy guidelines for apps used by minors and spurred the European Commission’s draft AI Act, which bans systems that exploit children’s vulnerabilities.

Translating the lessons to AI

Law‑makers say the same logic applies to generative AI tools that can produce endless personalised content for kids. The FTC’s recent “AI for Kids” notice asks developers to disclose how their models collect data and to implement age‑verification mechanisms. In the EU, the AI Act draft includes a “high‑risk” category for AI that targets children, mandating impact assessments and a ban on manipulative techniques.

What regulators are proposing now

• U.S. Federal Trade Commission: In a September 2024 statement, the FTC announced a rulemaking docket that would require AI developers to label content as “generated” and to obtain parental consent before collecting behavioural data from anyone under 13.
• European Commission: The AI Act, expected to be finalised in early 2025, calls for independent audits of child‑focused AI products and a mandatory “safety‑by‑design” clause.
• State‑level action: California’s new Consumer Privacy Act amendment, passed in November 2024, extends existing child‑privacy protections to AI‑driven apps.

These measures echo the “fair‑information‑practice” principles that emerged after the social‑media backlash, but they also try to move faster. Critics argue that the tech industry is already lobbying for loopholes, much as it did during the social‑media hearings.

Early wins and lingering doubts

Meta recently pledged $1 billion over three years to fund research on AI safety for children, echoing its earlier $1 billion commitment to child‑safety on its platforms. OpenAI, meanwhile, has introduced a “child‑mode” for its ChatGPT product, limiting access to certain topics and disabling data‑collection features for users under 13.

Yet many questions remain unanswered. The FTC has not clarified how it will verify compliance, and the European Commission’s enforcement framework is still vague. Moreover, no legislation yet defines what “manipulative” means in the context of AI‑generated content, leaving a potential loophole that tech firms could exploit.

The road ahead

If the social‑media reckoning taught anything, it is that waiting for a crisis to erupt is costly. Policymakers are now trying to act pre‑emptively, but the speed of AI development may outpace the slow grind of regulation. Observers say the next test will be whether regulators can enforce the rules they are drafting, or whether the industry will find new ways to sidestep them.

Bottom line: The same forces that finally forced social‑media platforms to answer for kids’ screen‑time are now being marshalled to keep generative AI from becoming the next unchecked addiction.