This is the story I keep pointing to when people ask whether age-verification laws actually protect kids, or just let platforms look like they're protecting kids. In January 2026, Roblox rolled out chat-access age checks in Australia, the kind of feature companies point to when regulators ask what they're doing about grooming and child safety. Then Australia's eSafety Commissioner tested the feature itself instead of taking Roblox's compliance claims at face value. Adults could still send connection requests directly to children, and kids' profiles remained broadly visible across the platform, despite the age checks supposedly being in place[1][2].

On Aug. 20, eSafety secured a court-enforceable undertaking from Roblox, not a fine, not a warning letter, an actual legally binding commitment. Roblox now has three months to implement real controls and submit to Australia's first-ever third-party age-assurance audit under the Online Safety Act. If it doesn't follow through, eSafety can go straight to the Federal Court for a compliance order[1][3].

The audit requirement is the part I think matters most. A lot of child-safety enforcement stops at the company saying it fixed the problem. This doesn't. An independent third party will actually check Roblox's age-assurance system instead of accepting the company's own word for it, which is exactly the verification gap that let a broken feature stay live and marketed as working in the first place.

I don't think this makes Roblox uniquely bad. I think it makes Roblox the platform that happened to get independently tested this time. The uncomfortable question is how many other platforms are running safety features right now that nobody outside the company has actually verified work as advertised.