Algorithmic curation: Social media design and addiction traits

A recommendation feed doesn't just decide what a teenager sees, it learns what keeps them watching and delivers more of exactly that. Algorithmic curation is one of the central design theories in MDL 3047, the litigation alleging that major social media platforms were built to prioritize engagement over the safety of young users. Atraxia Law evaluates whether this specific feature contributed to a documented condition in your child, and whether you're eligible to file a social media addiction claim.

Legal theory behind algorithmic curation

The core allegation is not that platforms recommend content, but that the optimization objective behind those recommendations can favor material that generates a strong emotional response, even when that material is harmful to a vulnerable user. A chronological feed simply shows what was posted most recently, while an engagement-oriented system continuously predicts what a specific user is most likely to watch next.

A 2024 study of over one million TikTok videos delivered to 91 participants found that algorithms belonging to users with eating disorders delivered 146% more appearance-focused videos and over 4,000% more content the researchers described as analogous to pro-anorexia material, compared to healthy controls. That gap existed even though the two groups reported only modest differences in how much they said they liked such content.

How the algorithmic curation mechanism works

Two technical choices sit at the center of the alleged defect. Neither is presented as harmful on its own, but together they're alleged to narrow a vulnerable user's feed around their own distress. Plaintiffs point to:

• Collaborative filtering and behavioral profiling, where signals like watch time, completions, and repeated views train the system to predict what a user will engage with next
• Amplification of emotionally intense content, since material that's distressing or triggering can generate the same engagement signals as content a user genuinely enjoys

Because the system interprets sustained attention as a request for more of the same, a single vulnerable moment can compound into a personalized stream of related material over time.

Alleged injuries linked to algorithmic curation

The litigation frames these harms as arising from repeated, personalized exposure rather than a single instance of encountering difficult content. Several overlapping pathways show up in research on affected populations. The harms alleged run through:

• Mental health deterioration that picks up pace via repeated exposure, social comparison, and reinforcement of pre-existing concerns
• Body image distortion, backed by research linking appearance-focused recommendations to increased upward appearance comparison
• Ideation loops, describing a self-reinforcing cycle where initial interest evolves into increasingly concentrated content over time

For 61 adolescents, exposure to self-harm content was associated with greater urges to self-injure during an eight-week 2025 study. A case evaluation looks at whether a documented diagnosis connects to this kind of exposure pattern on a specific platform.

Design alternatives platforms already had available

Part of the legal theory rests on the fact that less personalized systems already exist and were not applied by default to minor accounts. These aren't untested concepts, since some platforms already offer versions of them as opt-in settings. Documented mitigation approaches include:

• Chronological-only feeds by default, which reduce how much an algorithm determines what a user sees next
• Strict content guardrails for minor accounts, limiting recommendation and amplification of self-harm, eating disorder, and extreme dieting material without necessarily removing it from the platform
• Transparent recommendation logs, giving users and researchers visibility into why specific content was surfaced

Researchers studying adolescent self-harm exposure have specifically pointed to recommendation transparency as important for understanding risk. The existence of these alternatives doesn't prove a different design would have prevented any specific injury, but it supports the argument that engagement-driven personalization was a deliberate choice.

Talk to Atraxia Law about a social media addiction claim

If your child's mental health took a downturn following a pattern of algorithmically recommended content, Atraxia Law can look into whether the situation fits the current social media addiction litigation. We review the platform features involved together with your child's medical history to determine eligibility, then connect qualifying families with an attorney handling claims in MDL 3047. Reach out today to have your case evaluated at no cost.