The biggest step forward in safety since the invention of the internet
In London, some 300 workers in TikTok's trust and safety division have learned their roles will be absorbed by artificial intelligence, a quiet but consequential moment in the long negotiation between human judgment and algorithmic efficiency. The redundancies arrive precisely as the UK's Online Safety Act imposes new legal duties on platforms to shield children from harm — a paradox in which the law demanding more protection coincides with fewer human eyes watching. TikTok's parent ByteDance is wagering that machines can bear the moral weight once carried by people, even as regulators have yet to confirm that those machines are equal to the task.
- Around 300 London-based moderators and security staff received redundancy notices by email, learning their livelihoods had been quietly handed to an algorithm.
- The cuts land at the worst possible moment — just as the Online Safety Act activates, legally obligating platforms to do more to protect children, not less.
- Ofcom has not endorsed TikTok's AI age-assurance systems, leaving a visible gap between what the company is deploying and what the regulator considers adequate.
- Concern about the new restrictions has driven public interest in VPNs, prompting the Labour government to firmly rule out any ban while warning platforms not to encourage circumvention.
- TikTok is betting that artificial intelligence can deliver compliance more cheaply than human moderators — a wager that regulators will be scrutinising closely in the months ahead.
TikTok has begun notifying roughly 300 London-based employees in its trust and safety department that their roles are being eliminated, replaced by artificial intelligence systems designed to automate content moderation. The Communication Workers Union confirmed the notifications, which also affect trust and safety teams across South Asia and Southeast Asia as part of a wider ByteDance restructuring.
The timing carries a sharp irony. The redundancies coincide with the implementation of the UK's Online Safety Act, legislation that places strict new obligations on technology companies to identify harmful content and protect children from exploitation. Rather than growing its human workforce to meet those demands, TikTok is moving in the opposite direction, deploying AI-driven "age assurance" controls in place of trained moderators. Ofcom, the UK's media regulator, has not yet formally endorsed these systems, leaving an unresolved question about whether automation can meet the law's intent.
The legislation has also stirred public debate about virtual private networks, tools that could allow users to sidestep age verification requirements. The Labour government has rejected any suggestion of a VPN ban. Tech Secretary Peter Kyle described the Online Safety Act as the most significant advance in child safety since the internet's invention, and called on adults to cooperate with verification systems rather than circumvent them. He added that platforms found to be directing users toward workarounds could face legal consequences of their own.
What remains unresolved is whether algorithms can reliably do what human moderators were trained to do — recognise nuance, catch edge cases, and protect the most vulnerable users. TikTok is betting they can. Regulators will be watching to see if that bet holds.
TikTok has begun notifying its UK workforce of a significant restructuring that will eliminate roughly 300 jobs from its London-based moderation and security teams. The announcement came via email to staff in the company's trust and safety department, according to the Communication Workers Union, which represents affected workers. The cuts are part of a broader shift toward automating content moderation work through artificial intelligence, a move that will reshape how the platform manages the vast volume of user-generated content flowing across its service.
The timing of these redundancies is notable. They arrive just as the UK's Online Safety Act takes effect, new legislation designed to place strict legal obligations on technology companies to identify and remove harmful and illegal content while protecting children from exploitation and abuse. Rather than expanding its human moderation capacity to meet these heightened regulatory demands, TikTok is moving in the opposite direction—replacing people with algorithms. The company's parent firm, ByteDance, indicated that the restructuring will affect not only the London operation but also trust and safety teams across South Asia and Southeast Asia.
The shift toward automation reflects a broader industry trend, but it raises immediate questions about effectiveness. TikTok has introduced what it calls "age assurance" controls to comply with the Online Safety Act's requirements, systems that rely on artificial intelligence to verify user age and restrict content access accordingly. However, the UK's media regulator Ofcom has not yet formally endorsed these AI-based systems, leaving a gap between what the platform is deploying and what regulators have deemed acceptable.
The job cuts also come amid public concern about how the Online Safety Act will be enforced and what workarounds users might employ to circumvent new restrictions. The legislation's implementation has sparked increased interest in virtual private networks, or VPNs, tools that mask a user's location and identity online. Some observers speculated this might prompt the government to ban such services. The Labour government has firmly rejected that possibility. Tech Secretary Peter Kyle stated that the administration is "not considering a VPN ban," arguing instead that most people will comply with age verification requirements voluntarily. He framed the Online Safety Act as "the biggest step forward in safety since the invention of the internet" when it comes to protecting children, and urged adults to cooperate with verification systems rather than seek technical workarounds.
Kyle also warned that platforms which actively direct users toward VPNs or other circumvention methods could themselves face legal consequences under the new regulatory framework. The message is clear: the government expects both platforms and users to operate within the bounds of the law, not around them. Yet TikTok's decision to automate rather than hire suggests the company may be betting that artificial intelligence can handle compliance more cheaply and efficiently than human judgment ever could. Whether that bet pays off—whether algorithms can reliably identify harmful content and protect vulnerable users as effectively as trained moderators—remains an open question that regulators will be watching closely in the months ahead.
Notable Quotes
The biggest step forward in safety since the invention of the internet. When it comes to children, that is something we celebrate.— Tech Secretary Peter Kyle, on the Online Safety Act
If platforms or sites signpost towards workarounds like VPNs, then that itself is a crime and will be tackled by these codes.— Tech Secretary Peter Kyle, warning against circumvention of age verification
The Hearth Conversation Another angle on the story
Why is TikTok cutting jobs right now, of all times? Isn't this when they need more people to handle the new rules?
That's the paradox. The Online Safety Act creates legal pressure to moderate better, but it also creates financial pressure. TikTok is choosing to solve that with automation rather than hiring.
And they think AI can do what 300 people were doing?
They're betting on it. But here's the thing—Ofcom hasn't even approved their age assurance systems yet. They're moving fast, maybe faster than regulators are comfortable with.
What happens to the people losing their jobs?
The union expects around 300 redundancies in London alone. These are skilled workers in trust and safety—people trained to make judgment calls about what should and shouldn't be on the platform.
Can an algorithm make those calls?
That's what regulators are trying to figure out. Content moderation isn't just about rules; it's about context, nuance, cultural difference. AI is getting better, but it's not there yet.
So this could backfire?
If the platform fails to catch harmful content, or if it over-moderates and silences legitimate speech, yes. And the government has made clear it will hold platforms accountable.