California lawmakers move to rein in Big Tech on AI and social media

[Depot76]

California legislators are pushing forward a wave of bills aimed at regulating both artificial intelligence and social media platforms, driven by what one assemblyman described as a heightened level of tech anxiety spreading from job losses to data centers to AI itself. One notable bill, Assembly Bill 1709, would bar certain online platforms from offering addictive features like autoplay and algorithmically recommended feeds to users under 16, while also adding new age verification requirements. Governor Gavin Newsom, who has previously vetoed some broader tech restriction bills, will ultimately decide whether to sign this new batch into law

Newsom himself framed California's approach as learning from and working alongside European regulators rather than simply copying their model outright, saying the state intends to keep pushing boundaries and litigate where necessary. Researchers and advocates argue that autoplay and endless recommended feeds specifically are harmful to developing brains and shouldn't be considered appropriate design choices regardless of any broader AI debate. The tech industry, represented by groups like TechNet whose members include Meta and Snap, has pushed back against several of these measures as regulatory overreach

California has positioned itself as the state most willing to actually legislate ahead of federal action on AI and social media, partly because so much of the industry is headquartered there in the first place. Curious what people think about California specifically taking the lead on this rather than waiting for federal rules

Reining in Big Tech: How California lawmakers plan to regulate AI and social mediawww.latimes.com

[NatureBoyLewis42]

Not entirely sure how much of this genuinely protects users versus just generating headlines for legislators ahead of an election cycle. Not saying the concerns aren't real, just that political incentives and actual policy substance don't always line up neatly. Would want to see the final signed language before judging how meaningful this actually turns out to be

[Anvil79]

Age verification requirements always sound reasonable in theory but tend to get genuinely messy in actual implementation, raising real privacy concerns about who ends up holding sensitive identity data. Worth watching closely how California actually handles that specific technical challenge. Good intentions don't automatically translate into good execution here

[Rapid Ava]

Autoplay and algorithmic feeds specifically deserve the scrutiny they're finally getting honestly, since the addictive design choices behind them were never really about genuine user benefit in the first place. Glad to see lawmakers naming those specific mechanisms directly instead of vague AI language. Precision in the actual bill language matters a lot here

[Laura94]

California leading on this stuff makes sense given how much of the industry is headquartered there, but it also means state level rules end up functioning as de facto national policy for products used everywhere. Companies rarely build separate versions just for California specifically. Worth remembering these bills have national reach in practice regardless of state borders

[Sting]

Newsom's veto history on tech bills makes me genuinely skeptical this entire package survives fully intact once it actually lands on his desk.

He's shown a real pattern of balancing industry relationships against public pressure before. Wouldn't be shocked if some of these specific provisions get watered down or cut entirely

[Ederson]

The pace is probably the hardest part. A law can take a long time to draft and implement, while a software product can be redesigned in a few weeks. By the time a very specific rule takes effect, the technology it was written for may look completely different.

Rules based on outcomes and basic duties could age better than rules written around one particular model or platform feature. Otherwise lawmakers risk playing permanent catch-up.

[ArVeeDee]

One practical improvement would be making companies keep better records of how automated systems are tested and changed. If something goes wrong, investigators should be able to establish what system was being used, what safeguards existed and whether anyone noticed warning signs.

That feels more useful to me than trying to predict every possible AI failure in advance. Technology changes too quickly for legislation to list every future scenario.

[James93]

One thing that gets overlooked is that AI regulation and social media regulation aren't necessarily the same problem. An algorithm recommending videos, a generative AI system producing text and an automated hiring tool can have completely different risks.

A single sweeping framework could end up being either too vague to work or unnecessarily restrictive in areas that don't need the same treatment. Targeted rules may be less exciting politically, but they are easier to understand and test in practice.

[Pilot]

There is a legitimate free-speech question around social media rules too. Nobody wants illegal or genuinely harmful material treated as untouchable simply because it is online, but rules affecting moderation can have consequences for perfectly lawful speech as well.

That doesn't mean platforms should get a free pass. It means the wording matters enormously. A requirement to provide explanations or appeals is quite different from telling a private service exactly which opinions it must allow or remove.

[Sabu_ECW]

The difficult part is turning broad concerns about AI and social media into rules that are specific enough to enforce. Saying a platform should be "responsible" sounds good, but lawmakers still have to define what that means when a recommendation system, chatbot or moderation tool causes harm

That is where I think the details of each bill matter more than the overall headline. A narrowly written transparency requirement is a very different thing from a rule that effectively dictates how a product has to operate

[Tel86]

A lot of these debates seem to jump straight from "AI can cause harm" to "therefore regulate AI" without discussing what the regulation is supposed to achieve. Protecting children, reducing fraud, improving privacy and controlling automated decision-making are all different objectives

If a bill clearly identifies the problem and gives companies a measurable requirement, at least there is something meaningful to debate. If the requirement is basically "make the algorithm safer", lawyers are going to spend years arguing over what that means

[FinalDavid14]

Newsom's previous decisions are worth watching, but a bill reaching his desk is not the same thing as a final policy. There can be amendments, negotiations and completely different outcomes for different measures in the same package

I'd rather see the individual proposals judged on what they actually require. Bundling everything together as either "Big Tech regulation" or "government overreach" makes it almost impossible to have a useful discussion about the details

[RicFlair04]

The smaller companies deserve some attention here. A huge platform may have dedicated safety teams, auditors and engineers who can build compliance tools, while a startup with twenty employees might struggle with the same paperwork

If lawmakers want the rules to work across the industry, sensible thresholds and proportional requirements would make more sense than treating every company as though it has the same resources. Otherwise regulation can unintentionally become a moat protecting incumbents

[PolicyDrifter]

Whatever happens with this particular collection of bills, the bigger issue isn't going away. Companies will keep deploying more automated systems, and governments will keep trying to decide where existing laws are sufficient and where new safeguards are needed

The useful question is probably less "should California regulate AI?" and more "which specific harms are these rules trying to prevent, and can the proposed requirements actually do that?" That is where the interesting debate starts

[QuietObserver34]

The strongest case for action, in my view, is where an individual has very little ability to protect themselves. A child using a social platform or someone whose personal data is being processed by an automated system may not have the knowledge or bargaining power to negotiate better terms

That doesn't mean every AI feature needs a government rule attached to it. Consumer protection works best when it focuses on identifiable risks rather than treating the entire technology as inherently suspect

[Pilgrim11]

There is also a funny contradiction in some of these arguments: people want powerful AI companies to be controlled, but they also want rapid innovation and cheap access to the technology. Those goals can coexist, but not automatically

Good regulation should ideally make the dangerous shortcuts harder without making ordinary experimentation impossible. Getting that balance right is much harder than announcing that regulation is either good or bad.