Texas judge rules TikTok misled users on child safety

TL;DR: A Texas district judge ruled on 10 September 2026 that TikTok violated the Texas Deceptive Trade Practices Act by telling users it removed rule-breaking content when its internal policy instead marked some of it merely “hard to find”, and by promoting its Restricted Mode setting as a filter for minors while its own internal reviews showed the filter let through material it was supposed to block. Liability is now settled before trial — the trial, expected to be scheduled within weeks, will decide penalties of up to US$10,000 per violation and whether TikTok must change how its moderation actually operates. The ruling lands four months after Malaysia’s under-16 social media rules came into force, and it gives Malaysian parents one concrete reason to stop treating a single “restricted” toggle as proof that an app is safe.

Texas judge rules TikTok misled users on child safety is not a story about a fine, at least not yet. No penalties have been assessed, no app has been banned, and the case is not over. What has already happened is narrower and more interesting: a court has decided, before any jury hears the evidence, that TikTok’s public statements about how it protects children did not match what its own systems did. That kind of finding — deception established at the summary judgment stage — is the first of its kind against TikTok anywhere in the United States, and it turns a decade-old debate about “content moderation” into a much simpler question: when a platform says a video is removed, was it removed or was it just harder to find?

Here is what the judge actually decided, how the mechanism worked, what the remedies phase could cost, and what any of it means if you are reading this in Kuala Lumpur rather than Austin.

What the Texas Judge Ruled: TikTok Misled Users on Child Safety

The order came from Judge Cory Liu of the 250th District Court in Travis County — the Austin court that hears most cases brought by the State of Texas — and it granted the state’s motion for partial summary judgment. In plain terms, Texas asked the court to rule on liability without a trial, arguing that the undisputed evidence already showed TikTok had deceived consumers. Judge Liu agreed on two counts, in an order dated 10 September 2026 and widely reported the following day.

The case itself was filed on 9 January 2025 by Texas Attorney General Ken Paxton, who accused TikTok of deceptively marketing its app as safe for children despite what his office described as an endless stream of explicit and inappropriate material reaching minors. The petition alleged ten separate violations of the Texas Deceptive Trade Practices Act (DTPA), one of the state’s oldest consumer protection statutes, covering everything from alcohol, tobacco and drug references to sexual content, profanity, and the age rating the app carried in major app stores.

The two allegations the judge accepted are the ones that matter:

1. The “removal” claim. Texas showed that TikTok told users content violating its Community Guidelines would be removed, while internal moderation policy in some cases directed that material to remain available under the designation “hard to find” rather than “do not allow”. The court found that practice violated state consumer protection law.
2. The Restricted Mode claim. Texas argued that TikTok misled users when it said Restricted Mode effectively filtered content that could be inappropriate for minors. Internal communications acknowledged the setting did not work as intended and “exposed users to a high amount of content that should have been filtered out”, according to the filings. Judge Liu found the state had established that Restricted Mode did not work as advertised.

Paxton’s response was blunt. “TikTok sacrificed the safety and innocence of children for engagement and numbers, and now they are being held accountable,” he said in a statement, adding that Texas was the first state in the country to hold TikTok liable for misleading parents about the safety of its platform.

TikTok did not immediately respond to a request for comment made outside regular business hours. Practically, the ruling means the company can no longer argue at trial that it did not deceive anyone: that question has been answered, and only the consequences remain.

Why “Hard to Find” Is the Heart of the Case

Strip away the legal language and this is a story about a moderation taxonomy that nobody outside the company could see.

Most large platforms sort content decisions into a small number of outcomes. A video can be removed outright — gone for everyone, no link works. It can be demoted, meaning the recommendation engine stops pushing it but it remains searchable, linkable and reachable through a profile. It can be labelled, fact-checked or age-gated. Each of those outcomes produces a very different experience for a user, and only one of them matches the sentence “content that violates our Community Guidelines is removed”.

The categories named in the case are the ones every parent would recognise: videos promoting drug use or gambling, minors behaving in a sexually suggestive manner, graphic wounds or injuries, plus profanity, violence, weapons and other mature themes. The deception was not that these videos existed — objectionable content exists on every large platform — but that the platform’s own public descriptions of its filters implied a level of removal and exclusion the internal system was never designed to deliver.

That distinction is why “hard to find” became the phrase the whole case now turns on. It is an internal comfort word. To a trust-and-safety team, “hard to find” sounds responsible: the content is demoted, not recommended, not surfaced to minors on a For You feed. To a parent reading a Help Centre page that says violations are removed, it means nothing at all, because the parent has no way to discover that the label exists. When a public promise and an internal label diverge, the gap between them is where the legal risk lives — and in Texas, that gap has now been ruled a violation of consumer protection law.

What Happens Next: Trial, Penalties and a Possible Injunction

Because the ruling came by partial summary judgment, the case now moves to trial on remedies alone. Paxton’s office said the trial is expected to be scheduled within about a month, which points to October 2026 if the calendar holds. If no settlement is reached beforehand, it would be the first full trial over TikTok’s child safety claims — a rare event, since most platform liability cases so far have ended in settlements before full argument on the facts.

At that trial, the court is expected to work through three things: civil penalties, statutory damages, and permanent injunctive relief.

• Civil penalties. Under the DTPA’s enforcement provisions, the state can seek up to US$10,000 per violation. Texas has framed the misconduct as per-instance conduct rather than one continuous act, which is how a case touching millions of Texas users becomes a number with a lot of zeroes behind it. The state is seeking fines of up to that per-instance figure, not a fixed sum, so the total is a function of how the court counts violations.
• Statutory damages. Separate from the penalties, damages can be calculated against the same underlying misrepresentations if the court treats them as remediable under the statute.
• Injunction. This may end up being the most consequential outcome for users outside Texas as well. A permanent injunction could require changes to how TikTok classifies and discloses content decisions — the kind of order that forces a company to make its taxonomy legible to the public instead of keeping “hard to find” as an internal-only label.

Texas is not running this case in isolation. The same office sued TikTok in October 2024 over the state’s Securing Children Online Through Parental Empowerment (SCOPE) Act, arguing the platform failed to provide parents with required tools and safeguards for known-minor accounts; that case seeks civil penalties of up to US$10,000 per violation plus injunctive relief of its own. The attorney general’s office has applied a similar template elsewhere, pairing DTPA claims with nuisance claims against platforms accused of facilitating harm to minors. In other words, Texas has assembled a portfolio of child safety cases and now has a liability finding to point at in all of them.

There is also a straightforward commercial reason to expect a settlement rather than a courtroom. In August 2026, TikTok agreed to settle three US lawsuits brought by young people who accused social media companies of designing addictive platforms that harmed their mental health. The company has shown it will pay to close litigation when the exposure becomes unmanageable — and this case has now moved into precisely that phase.

Meta’s $18 Billion Deal Is the Benchmark TikTok Is Being Measured Against

To understand the pressure on TikTok, look at the settlement its biggest rival signed weeks before Judge Liu’s order.

In August 2026, Meta agreed to pay up to US$18 billion to resolve claims from a bipartisan group of 52 attorneys general across US states, territories and the District of Columbia that it had misled the public about the safety of Facebook and Instagram for young users. A judge approved the agreement later in the month. Under its terms:

• Meta pays roughly 70 percent of the total — about US$12.7 billion — in annual instalments over ten years.
• The remaining 30 percent, around US$5.3 billion, is released only if YouTube and TikTok adopt matching measures: a one-hour daily limit for teens, a Night Mode block, and age assurance systems — and each pays an amount matching part of that remaining sum.
• Instagram and Facebook carry a default two-hour daily limit for under-18s that only a parent can lift, a default midnight-to-6am blackout, prompts after 15 minutes of continuous use, an independent auditor, and an injunction barring further false or misleading claims about its safety features.
• If rivals join the framework, the commitments strengthen — the default limit drops to one hour and the night block extends to 10pm–7am.

Meta publicly called on TikTok and YouTube to adopt what it described as a new industry standard. Neither company has rushed to sign on, and Meta’s own conditional structure means the announcement doubles as competitive pressure: the settlement explicitly names rival platforms and makes part of Meta’s payment contingent on their behaviour.

The contrast with Texas is instructive. Meta bought peace across most of the country; TikTok, in Texas at least, has a court finding that its safety messaging was deceptive and no deal to point to. And Texas is not the only state pursuing TikTok — California, New York and a group of other attorneys general have their own ongoing consumer protection case against the platform, with California’s attorney general stating publicly that he expects TikTok to adopt practices similar to Meta’s commitments. Meanwhile, separate litigation against Meta in New Mexico produced penalties of more than US$900 million after a court found child safety law violations.

The direction of travel is clear: states have stopped waiting for federal legislation and have started obtaining enforceable code-of-conduct style commitments through courts and settlements. Platforms that cannot show their safety features measurably work are the ones in the line of fire.

What the Texas Ruling Means for Malaysian Parents and Users

The Texas case has no direct legal effect in Malaysia — it is a state consumer protection action under Texas law. But it arrives at a moment when Malaysian rules demand almost exactly the thing Texas has just found TikTok failed to provide: safeguards that actually function, not safeguards that exist as toggles and labels.

Malaysia’s Online Safety Act 2025 (ONSA) framework changed the ground rules in 2026. The Child Protection Code (CPC) and the Risk Mitigation Code (RMC) took effect on 1 June 2026, and they apply to licensed social media platforms with at least eight million users in the country — a list that includes TikTok alongside Facebook, Instagram and YouTube. The core requirements:

• Minimum age 16. Users below 16 are not permitted to register social media accounts, and platforms must ensure only those aged 16 and above can register and access age-appropriate features.
• Government-backed age verification. Verification must be done against government-issued records or equivalent recognised documents — MyKad, passport or MyDigital ID rather than self-declaration. Existing accounts are being verified progressively over a period of up to six months, which puts full coverage around the end of 2026.
• Safety by design and real moderation. Under the RMC, platforms must run risk assessments, operate content governance and moderation, adjust recommendation systems to reduce exposure to harmful material, provide reporting and response mechanisms, verify advertisers, and label manipulated content where appropriate.
• Real penalties. Non-compliance can attract financial penalties of up to RM10 million.

Read those obligations next to Judge Liu’s findings and the overlap is uncomfortable. Malaysian rules require platforms to implement effective, appropriate and compliant safeguards — MCMC has been explicit that it is technology-neutral about how age verification is done, but not about whether it works. A platform that advertises a restricted setting while internally treating it as a soft demotion is precisely the pattern the Texas court has now labelled deceptive. In Malaysia, the same gap between the promise and the enforcement action is a compliance failure with a statutory fine attached, no court finding required.

For parents, the practical lesson from Texas is simple: audit the feature, do not trust the label.

1. Test Restricted Mode yourself. Turn it on, then search for the categories it claims to filter. If videos that should be blocked still appear, you have your answer — and you should not rely on that setting alone.
2. Use the supervision tools that actually enforce limits. Family Pairing on TikTok allows a parent to link accounts, set screen-time limits, restrict direct messages, filter keywords and control who can comment. Those controls act on the account, not on a mood.
3. Do not treat the 16+ rule as done. Age verification is still rolling out across existing accounts in Malaysia. If your child is under 16, the law says they should not have their own account — and if they use yours, supervision and screen-time limits are the practical substitute.
4. Keep identity documents out of the wrong pipeline. Malaysia’s rules embed verification in government-issued records; use the platform’s official in-app verification flow and never send MyKad images to third parties offering to “verify” an account for you.
5. Report, in bulk if needed. Content moderation systems respond to volume. Reporting a video does double duty: it removes that video and it creates a record that regulators can measure a platform against.

For Developers: The Moderation-Taxonomy Lesson, in Code

If you build any product with user-generated content — a community app, a marketplace, a comment system — the Texas case is a design brief about honesty in your own data model.

The failure was not a missing feature. It was a mismatch between the enforcement action your system takes and the words you use to describe that action to the public. That is testable in code, and the test is worth writing before a regulator writes it for you.

# promise_vs_enforcement.py
# The check Texas effectively ran against TikTok's public statements:
# does every policy outcome map to the action users were promised?

PUBLIC_PROMISE = {
    "guideline_violation": "removed",          # what the Help Centre said
    "minor_inappropriate": "filtered_in_restricted_mode",
}

# What the enforcement layer actually did.
ENFORCEMENT = {
    "remove":         "content unavailable to everyone",
    "hard_to_find":   "demoted in recommendations, still searchable",
    "informational":  "labelled, still fully in feed",
}

POLICY_MAP = {
    "guideline_violation": "hard_to_find",     # <-- the gap
    "minor_inappropriate": "informational",
    "spam":                "remove",
}


def audit(policy_map=POLICY_MAP):
    """Return every promise the product cannot honour."""
    return {
        promise: {"promised": action, "delivered": policy_map.get(promise, "unset")}
        for promise, action in PUBLIC_PROMISE.items()
        if policy_map.get(promise) != action
    }


if __name__ == "__main__":
    gaps = audit()
    for promise, detail in gaps.items():
        print(f"MISMATCH {promise}: promised={detail['promised']} "
              f"actual={detail['delivered']}")
    # -> two mismatches. In a court, each one is a representation to consumers.

If that function returns more than an empty dictionary, you have two options: fix the enforcement action so it matches the promise, or rewrite the promise so a reader can verify it. What you cannot do is keep the comfortable label and hope nobody measures it. Texas measured it, and the comfortable label — “hard to find” — became the evidence.

FAQ: The Texas TikTok Child Safety Ruling

What exactly did the judge decide?
Judge Cory Liu of the 250th District Court in Travis County granted Texas partial summary judgment on two claims under the Texas Deceptive Trade Practices Act: that TikTok misled users by stating violating content was removed when internal policy designated some of it only “hard to find”, and that it misled users about the effectiveness of Restricted Mode for minors. Liability on those two points is established before trial.

Did TikTok lose the whole case?
No. This is a partial summary judgment on liability, not a final judgment on money or remedies. The case proceeds to trial, expected to be scheduled within about a month, where penalties, statutory damages and possible injunctions — orders requiring changes to how the platform works — will be decided. TikTok can still contest the remedy phase, and a settlement before trial remains possible.

How much could TikTok have to pay?
Texas is seeking civil penalties of up to US$10,000 per violation under the state’s consumer protection law, framed per instance rather than as one event, which is how a case involving millions of users scales. No total has been set, and the court has not assessed any amount. The separate Texas case over the SCOPE Act seeks penalties at the same per-violation ceiling plus injunctive relief.

Is this related to Meta’s US$18 billion settlement?
They are different cases, but they are part of the same wave. Meta agreed in August 2026 to pay up to US$18 billion to settle claims from 52 attorneys general, with around US$5.3 billion conditional on TikTok and YouTube matching teen time limits, night-time blocks and age assurance. That framework — negotiated, not legislated — is now the benchmark states are measuring other platforms against.

Does the ruling change anything for users in Malaysia?
Not directly; it is a Texas state case. But Malaysia’s Online Safety Act 2025 framework has required licensed platforms, including TikTok, to implement under-16 age verification, safety-by-design measures, content governance and algorithm adjustments since 1 June 2026, with penalties of up to RM10 million for non-compliance. MCMC has said it expects age verification and safeguards to be effective, not just present — which is the same standard the Texas court has now applied, in a different legal language.

Should I still use Restricted Mode for my child’s device?
Use it, but do not assume it is a filter in the strict sense. Turn it on and test it with searches for the categories it claims to block. Pair it with Family Pairing controls, screen-time limits and, for under-16s in Malaysia, the requirement that they not hold their own account at all.

Conclusion: A Label Is Not a Safeguard

The Texas ruling against TikTok does not ban the app, shut down its servers or hand parents a magic setting. What it does is establish something harder to undo: that describing a demotion as a removal is a deception, and that calling a filter a filter does not make it one. Judges, not marketing departments, get to decide what those words mean when a few million people rely on them.

Three things follow from here. The remedies trial will set the price of the mismatch, and an injunction could force TikTok to describe its content decisions in terms users can actually verify. Malaysia’s ONSA regime will keep testing the same question from a different direction, since the CPC and RMC require safeguards that are effective and compliant, with fines up to RM10 million available for platforms that fall short. And for everyone else building software, the lesson travels further than one app: if your internal taxonomy and your public promise disagree, you have a defect — and unlike most defects, this one is measurable in a line of code.

For parents, the takeaway is unglamorous and reliable. Check the feature, not the label. Test Restricted Mode, use supervision tools that enforce limits at the account level, and assume that any platform’s safety claim is worth exactly as much as your own five-minute test of it.

More from ZFRBuild

• November 8 Diwali: The holiday every Malaysian should know
• Best robot vacuums under RM1,000 in Malaysia (2026)
• LG smart TVs caught logging audio with the screen off