New KGI Report Finds US Litigation Reveals Gaps in How Snap and YouTube Measure and Publicly Disclose Risks to Minors

Washington, D.C. – As social media companies face mounting legal challenges in the United States and increasing regulatory scrutiny in Europe, a new report from KGI finds a significant gap between what Snap and YouTube publicly disclose in their EU risk assessments and what has emerged through US litigation, including their own internal data, research, communications, and risk evaluation.

KGI’s latest report in the Measuring Risk series, Measuring Risk II: What EU Risk Assessments and US Litigation Reveal About Snap and YouTube, compares Snap and YouTube’s 2025 Digital Services Act (DSA) risk assessments with evidence emerging through US litigation, including internal company documents, depositions, and expert reports. The report focuses on risks related to the protection of minors and examines whether platforms provide meaningful evidence that their safety measures actually reduce risk.

“US litigation has opened an unprecedented window into how social media companies understand their own risks – from the product design features they build to the metrics they track internally,” said Peter Chapman, Associate Director at the Knight-Georgetown Institute. “The evidence shows that platforms have sophisticated capabilities to measure risk and evaluate mitigations, but those capabilities are rarely reflected in the public disclosures that regulators, researchers, parents, and users rely on to make informed technology choices.”

The report finds a consistent pattern across Snap and YouTube: both companies possess comprehensive internal data and robust analytical capabilities to measure user behavior, assess potential risks, and evaluate mitigation efforts, yet current EU risk assessments largely describe policies, features, and processes without reporting the outcome-based metrics needed to determine whether those measures are effective.

Key findings from the report include:

• Platforms possess robust measurement capabilities that could be used to assess risk and mitigation effectiveness. Across both Snap and YouTube, litigation records reveal collection of user behavior data related to risk. Internal strategies and studies at both companies reference a reliance on behavioral trace data and self-reported user experiences to understand user engagement patterns, late-night use, sleep disruption, and regret. Very little of this measurement capability is reflected in published EU risk assessments.
• Platforms promote safety features as key mitigations without demonstrating that they actually work. Snap and YouTube describe a range of safety interventions in their EU risk assessments – including age assurance, parental controls, screentime tools, and recommender system safeguards – yet neither discloses the adoption, usage, or outcome data needed to determine whether these interventions are effective. Snap’s flagship safety tool, Family Center, is cited more than 40 times in its risk assessment, but litigation documents indicate only 0.33% of teenage users were enrolled. Meanwhile, internal documents indicate that awareness of YouTube’s parental control tool, Family Link, was reportedly just 15% among parents in 2022.
• Internal records reveal significant gaps in the evaluation of key safety risks and mitigations. Snap’s CEO testified in deposition that the company employed no “full-time mental health researchers,” while its head of user research testified in deposition that the company had never conducted research on user experiences related to mental health, addiction, or anxiety. Snap’s former Lead Safety Product Manager testified in deposition that mental health work was “volunteer work,” as the company had no dedicated product manager focused on mental health or wellbeing. YouTube’s risk assessment lists multiple safety interventions, but includes scant details about effectiveness. Internal teams acknowledged in 2025 that they “don’t know if” Take a Break and bedtime reminders “work” because they had not looked “too deeply” into their effectiveness. 
• Platforms study engagement-based design risks internally but rarely disclose them in their public assessments. Litigation documents show both companies internally discuss and study risks associated with engagement-maximizing product designs, including gamification, notifications, and autoplay. Yet, public risk assessments continue to focus primarily on content- and user-generated risks, providing comparatively little evidence about the risks and effectiveness of platform design.

Measuring Risk II: What EU Risk Assessments and US Litigation Reveal About Snap and YouTube is the second report in KGI’s Measuring Risk series. The first report examined Meta and TikTok’s risk assessments and internal records emerging from US litigation. 

Together, the two reports show a consistent pattern across major social media companies: platforms already collect, test, and analyze data needed to assess whether mitigations are reasonable, proportionate, and effective, but rarely disclose the outcome-based metrics needed to demonstrate that those measures work in practice.

“Across Meta, TikTok, Snap, and YouTube, we found broadly consistent approaches to identifying and mitigating risk,” said Chapman. “Documents released in litigation clearly demonstrate that large social media platforms have the tools to ensure their designs and safety mitigations actually reduce risk. What’s missing isn’t capability, it’s a requirement to consistently and publicly prove effectiveness.”

Read the full report here.