A third law, a fourth deadline
On August 31, 2026, the California legislature passed Assembly Bill 1709, sending it to Governor Newsom's desk with a decision due by September 30. The bill, authored by Assemblymember Josh Lowenthal and co-authored by Josh Hoover, would bar "covered platforms" from offering algorithmically recommended feeds, autoplay, and other features the Attorney General designates as "addictive" to users under 16, plus a broader set of core social functions such as following and liking. Platforms could still let minors hold accounts, but only stripped of the features that make a modern feed a feed. The restriction takes effect January 2027 — the same month two other California youth-safety laws also come fully online (EFF, Aug 31, 2026; AB 1709 bill text).
That collision is the real story. California already has SB 976, the "Protecting Our Kids from Social Media Addiction Act," which Newsom signed in September 2024 and which bars addictive feeds for minors absent verifiable parental consent — its implementing regulations, including age-assurance standards, are due from the Attorney General by January 1, 2027 (oag.ca.gov). It also has AB 1043, the Digital Age Assurance Act, signed in October 2025, which builds a device-level age-signal system: operating systems ask for a birth date at setup and pass an age bracket (under 13, 13–15, 16–17, 18+) to apps via API, explicitly to avoid requiring ID uploads or biometric scans. AB 1043 becomes operative January 1, 2027 as well (AB 1043 bill text).
AB 1709 leans on AB 1043's signal infrastructure but, per EFF's July 2026 analysis, "introduces legal confusion by creating provisions that conflict" with both existing statutes — and because AB 1043 doesn't specify how a platform should resolve ambiguous or contested signals, AB 1709 will in practice push platforms toward stricter verification than either prior law contemplated (EFF, Jul 31, 2026). Three statutes, three compliance regimes, one age bracket, and no clarity yet on which controls when they diverge.
The case for the bill, stated fairly
Proponents aren't wrong that feed design matters. Common Sense Media, backing AB 1709, argues the bill targets "the design, not the speech" — platforms have "studied how kids behave, when they are most vulnerable, and how to keep them coming back," and handing a 14-year-old an engagement-optimized feed is closer to handing them "the digital equivalent of a pack of cigarettes" than to protecting their expression (Common Sense Media). Pew's most recent survey of US teens found 90% use YouTube, roughly 60% use TikTok and Instagram, and nearly half say they're online "almost constantly" — up from 24% a decade ago (Pew Research, Dec 2024). A bill that targets the recommendation engine rather than banning speech outright is, on paper, more narrowly tailored than a blunt access ban. That's a real distinction, and it's the reason AB 1709 cleared the Assembly floor 76-0 in May before advancing through the Senate.
Why the design still fails
The distinction collapses on contact with what a "feed" actually is. Stripping "addictive features" as AB 1709 defines them — recommendation algorithms, autoplay, and whatever the Attorney General adds by future rulemaking — doesn't leave a safer version of Instagram or YouTube for a 15-year-old. It leaves a reverse-chronological list with no discovery mechanism, which is a materially different, degraded product that happens to still carry the same brand name. EFF's argument that this restricts "virtually every functional recommendation algorithm and basic input" understates nothing: following and liking, which the news hook confirms are also barred, are not addictive dark patterns — they're how any social platform, including a public agency's community page or a mutual-aid forum, actually functions (EFF, Aug 31, 2026).
The deeper problem is regulatory design, not just First Amendment doctrine. California has now built three separate statutory age-verification regimes in three years, each with its own definitions of "minor," its own enforcement agency, and its own effective date, all converging on January 2027. AB 1043 was engineered specifically to let a teenager get an age signal without surrendering a government ID or a face scan to every app they open. AB 1709, layered on top without harmonizing definitions, risks forcing platforms back toward the harder verification methods AB 1043 was designed to make unnecessary — the exact "honeypot of sensitive personal data" EFF warns creates its own breach and surveillance risk (EFF, Jul 31, 2026).
There is also a straightforward litigation cost. NetChoice has already sued California over SB 976's feed restrictions; a third overlapping statute gives challengers a cleaner argument that the state's own laws are internally inconsistent, which tends to be fatal on vagueness grounds even before reaching the speech question. A legislature serious about protecting minors online has a simpler path available: let AB 1043's age-signal infrastructure and SB 976's consent-based feed restrictions actually take effect and be evaluated on results before adding a third, conflicting layer that no platform can cleanly implement and no court can cleanly parse. Newsom should veto AB 1709 not because youth-design harms are imaginary, but because California has already legislated an answer to this problem twice — and hasn't let either answer run.