Ofcom issued four penalties totalling £2.17 million in the period covered by its June 2026 online safety bulletin. The largest, £950,000, went to a suicide forum provider. 4chan was fined £540,000 for failing to carry out an adequate illegal content risk assessment, with daily penalties of up to £800 until it complies.
Five new investigations opened alongside them. Two of those, Teen Chat and Chat Avenue, are conversation platforms, examined over grooming risk and exposure to pornography. Telegram is under investigation over child sexual abuse material.
For anyone building or marketing a live social product in Europe, the significance is not the sums. It is that enforcement has arrived at the category, and the platforms named are chat services rather than the large feed-based networks everyone expected regulators to pursue first.
The Compliance Floor Moved
Two deadlines sit immediately ahead. Risk assessment records were due from specified providers on 31 July 2026, and the notification window for 2027/28 online safety fees closes on 30 September. Ofcom also publishes its categorisation register this summer, which determines which services fall into scope at all.
The practical effect is that safety architecture stopped being a differentiator and became a condition of operating. A platform without documented risk assessment, working age assurance and a moderation process it can evidence is not competing badly. It is exposed.
That is a meaningful shift for product teams who spent the last decade treating trust and safety as a cost centre attached to growth.
What Consent-First Design Actually Involves
The term gets used loosely, so it is worth being concrete. In live video specifically it tends to mean four things:
- Video blurred or off by default, with visibility requiring action from both sides rather than one
- Age assurance applied before access rather than after a report
- Moderation that scans sessions as they happen instead of waiting for a user to complain
- Exit and blocking controls positioned inside the interaction, not buried in settings
None of that is technically difficult. What makes it rare is that each item costs conversion. A platform that requires mutual consent before video starts will show worse funnel numbers than one that opens both cameras automatically, and for years the funnel won that argument.
The Ofcom penalties change the calculation, because the conversion gained by skipping those steps is now weighed against a fine plus daily accruals plus the categorisation register listing you as in scope.
Why This Became a Product Decision
Compliance-led design produces a recognisable outcome: safety features bolted on where an auditor will find them, and nowhere else. Users can tell the difference, and so increasingly can regulators, since Ofcom’s stated concern with 4chan was the absence of an adequate risk assessment rather than any single piece of content.
The platforms handling this well appear to have concluded that the safety layer is the product experience rather than a wrapper around it. Someone deciding whether to join a live girl video call or any other consumer video format is making a judgement about exposure in the first few seconds, well before they evaluate anything else the service offers. Whatever answers that question is doing more commercial work than the feature list.
There is a marketing consequence that tends to get missed. Trust signals have to be legible at the point of decision, not documented on a policy page. An icon showing that video starts blurred communicates more than a paragraph explaining a commitment to user safety, and it does so to the only audience that matters, which is the person deciding whether to continue.
Where the Category Goes Next
Two code amendments are expected in autumn 2026, covering intimate image abuse detection and crisis response. Both point in the same direction, toward proactive obligations rather than reactive takedown.
For operators, the planning assumption should be that the floor keeps rising and that retrofitting gets more expensive each time it does. For marketers, the more useful observation is that safety has quietly become a positioning axis in a category that used to compete purely on how quickly it could put two people in front of each other.
The platforms that treated that as a constraint are the ones now spending money on remediation. The ones that treated it as a design brief are the ones with nothing to remediate.
