Per‑child liability for platform harms

California authorised civil penalties up to $1 million per child against large social‑media companies found negligent in harming children, establishing a concrete per‑victim (per‑child) damages approach for online harms. The package also bans algorithmic 'addictive feeds' for under‑16s and requires pre‑deployment risk assessments for AI chatbots and state rules for independent AI auditors. — Imposing per‑child financial liability and mandatory AI auditor registries is a new regulatory lever that could reshape platform design, liability exposure and the economics of content moderation and AI safety on a national scale.