Internet Freedom Weekly: News Recap, July 6-10, 2026

Governments and platforms spent this week building systems that look like enforcement without doing any of the actual enforcing. For example, Australia's under-16 social media ban asked fifty test accounts to prove they were sixteen, and nine platforms took their word for it without asking twice. Meanwhile, the EU dusted off a chat control bill Parliament already voted down in March and revived it through a procedural rule that needs 361 votes to stop it rather than a majority to pass it. 

Between July 6 and July 10, that pattern played out across seven stories spanning Brussels, Canberra, Moscow, and New Delhi, with one EU vote still to come this week and every government insisting its system already works exactly as intended.

Missouri Joins the Age Verification Wave as SCOTUS Backs Texas App Store Checks

Missouri Governor Mike Kehoe signed HB 1839 on July 9, making age verification for adult sites permanent state law, with fines up to $10,000 a day for skipping it and up to $250,000 more if a minor gets through anyway. Pornhub had already blocked Missouri users rather than comply, joining more than 20 other states where it has pulled out entirely.

That same week, the Supreme Court declined to block Texas's App Store Accountability Act, letting the state keep requiring app stores to verify ages and secure parental consent before a minor can download anything. Missouri's law lets a platform simply exit the state to dodge compliance, but Texas moved verification up to the operating system layer, where leaving isn't an option.

Read the full article

Language Is the Internet Freedom Problem Nobody Discusses

Content moderation systems were built with English as the default, and Myanmar showed what that costs. When Facebook had almost no Burmese-speaking moderators during the mid-2010s, a UN fact-finding mission later concluded the platform played a "determining role" in violence against the Rohingya.

That same gap now shapes AI. Models trained mostly on English data perform worse in Yoruba, Quechua, or Tibetan, and UNESCO counts over 7,000 languages the internet barely serves. Kiswahili got an AI dictionary this week for World Kiswahili Language Day, a real step, but most languages aren't even close to that.

Read the full article

EU Finds a Procedural Trick to Force Chat Control Through Without a Real Vote

The European Parliament voted down extending chat control in March, 311 against to 228 in favor. The EPP revived the identical text anyway, routing it through the ordinary legislative procedure, which now needs 361 MEPs, an absolute majority, to block it instead of a simple majority to pass it.

A procedural vote lands Tuesday, with the real vote on Thursday, timed for the last plenary day before summer recess. Independent MEP Martin Sonneborn tried to block the maneuver by citing Parliament's own rulebook. Losing once apparently just means finding a rule that makes losing harder the second time.

Read the full article

Chat Control Won't Die: Inside the EU's Latest Push

Chat Control would have messaging apps scan private messages before encryption, catching known abuse images and grooming conversations alike. The proposal has cycled through the EU since 2022, and its "temporary" exemption from privacy law keeps getting renewed instead of expiring, most recently through August 2027.

Detection tools flag family photos and medical content along with real abuse material, and a false positive can trigger a police report before a human ever looks. Once a government builds a tool to scan everyone's messages for one category of content, that infrastructure doesn't stay limited to the category it started with.

Read the full article

Governments Don't Delete Posts Anymore. They Ban Outlets

OONI's Russia research found 279 news media domains blocked at the network level, double the 139 confirmed the year before, with outlets including Meduza and TV Rain unreachable from a standard Russian connection. When the EU blocked RT and Sputnik, Russia retaliated by blocking 81 EU outlets in return.

The Committee to Protect Journalists recorded 132 journalists and media workers killed in 2025 and 336 imprisoned, while Freedom House tracked people arrested for online expression in a record 57 of 72 countries. Block the newsroom, jail the journalist, and the message stops being about one article.

Read the full article

Australia's Under-16 Social Media Ban Can't Even Clear Its First Hurdle

Testers created 50 trial accounts across nine platforms covered by Australia's under-16 ban, each declaring the user was sixteen, and not one platform asked for proof. Some accounts received ads targeted at young audiences, and one registered on X was served adult content anyway.

Kick, a smaller live-streaming platform with less user data to lean on, was the only one of ten covered services that verified age before signup, while nine platforms with more resources simply estimated and let it slide. Australia has already deleted 4.7 million suspected minor accounts, and none of that paperwork proves anyone checked an age at the door.

Read the full article

How Digital ID Systems Make Refusal Impossible

India's Aadhaar launched in 2009 as a voluntary biometric ID and enrolled over a billion people before courts ruled mandatory linkage unconstitutional in 2018. By then it was already woven into banking, tax records, and welfare, and breaches of databases holding Aadhaar data have leaked hundreds of millions of records.

The EU's Digital Identity Wallet, rolling out under eIDAS 2.0, carries the same voluntary label Aadhaar once did. A system with 99.9% accuracy still fails a million people out of a billion, and those failures land hardest on people with the least power to contest them.

Read the full article