What the TikTok fine and the Gates interview tell us about platform moderation regulation

The date to fix on is 24 September. That is when TikTok withdrew its appeal against a £12.7m penalty issued by the Information Commissioner's Office in 2023, as The Guardian reported that day. The video app has a UK user base in the tens of millions. The fine was for inadequate checks on who was using the platform and for failing to remove underage users. When the ICO issued it, the regulator said TikTok had done "very little, if anything" to address underage use.

Three days later, on 27 September, Bill Gates went on NBC's Meet the Press and asked US legislators and law enforcement to regulate artificial intelligence. The Microsoft co-founder told host Kristen Welker that unchecked AI could cause "a billion deaths" and that "no one thinks self-regulation is enough".

Two stories, two continents, one argument underneath. Platforms and AI systems are not policing themselves. The bodies that are supposed to make them are moving at different speeds.

What TikTok actually gave up

The £12.7m sum is not new. What is new is that TikTok stopped fighting it. According to The Guardian's 24 September report, the company said it disagreed with the ruling but chose to pay because the case covered the period between 2018 and 2020, after which it had implemented "many youth safety measures and policies". That is a standard corporate formulation: agree to the bill, dispute the premise.

The second appeal matters more. TikTok also withdrew its challenge to an ICO request for information about how its algorithm processes teenagers' personal data and how it serves potentially harmful content. The ICO said it would now resume that investigation. That is an open file, not a closed one.

Emily Keaney, the ICO's deputy commissioner, framed the outcome in the regulator's usual terms: "Our successful action against TikTok shows how important it is for companies to have the right protections in place, including to prevent access by children who are not old enough to be using their services."

The numbers behind the case have not been seriously disputed. The ICO said up to 1.75 million UK children under 13 used the service in 2020. TikTok's own minimum age is 13.

The next deadline is bigger

Age assurance is about to become a much larger compliance problem in the UK. The government's under-16 social media limit comes into force next year, and Ofcom has said "more action from the tech industry is needed" on enforcing age limits as the ban approaches. TikTok's fine is, in that light, a rehearsal. It concerns a period that ended six years ago and a regime that did not yet exist.

The international evidence is not encouraging for anyone expecting clean enforcement. Australia introduced a social media ban for under-16s last December. The country's eSafety commissioner has said a "substantial proportion" of under-16s there have kept their accounts, created new ones or bypassed age-gating systems. A ban on paper is one thing. A ban at the login screen is another.

That gap between rule and enforcement is exactly the gap Gates was describing, in a different domain.

Gates wants a rulebook, not a kill switch

Gates's interview, aired on 27 September, was circulated in excerpt form in advance by NBC. He told Welker: "You need law enforcement and the politicians to get into the discussion about what safeguards and monitoring look like. And that has to be a required thing." He described the cost to industry as "a little bit of overhead", not a dramatic slowdown.

He was dismissive of the idea that a technical kill switch would be sufficient. "It's not enough to have a kill switch," he said, adding that current systems are not yet at the point where they "autonomously grab computers and, you know, can't be shut down". His broader warning, that AI in the wrong hands is "certainly powerful enough to drive events that, you know, cause a billion deaths", is the kind of claim that gets headlines and resists verification.

Context matters here. Gates published a nearly 6,000-word essay in late August warning about large-scale AI risks. The interview was one of his first major public appearances since becoming caught up in the Jeffrey Epstein affair. An external review commissioned by his foundation and released in July found that he was among staffers and leaders who collectively held about 30 meetings with Epstein between 2011 and 2014. Gates has not been accused of involvement in Epstein's admitted or alleged crimes, and he has said he regretted "every minute" he spent with him. He remains chair of the $89bn Gates Foundation.

Not everyone making the same argument has the same track record. That does not make the argument wrong. It does mean the messenger is part of the story.

One senator, one deadline, and a company under scrutiny

Gates is not alone in calling for a pause. Progressive US senator Bernie Sanders proposed legislation on 23 September that would temporarily pause advanced AI development while guardrails are put in place, The Guardian reported. Sam Altman and Elon Musk have also called for slower development, though the three men plainly mean different things by it.

There is an awkward subplot. Microsoft, the company Gates co-founded, has itself been challenged over its role in the systems driving some of these concerns. Employees and activists have organised a campaign to get Microsoft to cut ties with the Israeli military, which according to a Guardian essay published on Friday has used AI tools in Gaza and, alongside US forces, in Iran. The essay's authors, Mohamed Hussein and Granate Kim of the No Azure for Apartheid group, also wrote that the US Department of Homeland Security has used AI tools including Microsoft's Azure to monitor, arrest and deport immigrants. Those are allegations made in an opinion essay, not findings by a regulator.

They are nonetheless the kind of claims that make voluntary codes look thin. If the same company is both warning about AI risk and selling the infrastructure, self-regulation has an obvious structural problem.

Where the money is going

The market is already pricing the risk. On 24 September, Tech.eu reported that Kontext, an AI security startup founded by Jens Ernstberger and Michel Osswald, raised $4m in a round led by 42CAP with backing from a16z CSX and HTGF. Kontext builds a runtime security platform that sits between AI agents and the tools they use, evaluating each action against policy and blocking unauthorised ones. Teams can run it in observe mode first, then switch on enforcement.

Read that against the Gates interview and the shape of the coming regulatory market becomes visible. Regulators want auditable control points. Vendors are building them for AI agents before the rules exist. The same logic applies to age assurance: TikTok's fine, the UK's under-16 limit and Australia's experience all point at verification infrastructure rather than content takedowns.

Coverage of the TikTok case and the Gates interview diverges from the everyday reality of enforcement. The Guardian reported the ICO's action and the Australian eSafety commissioner's warning in the same piece. The two accounts do not contradict each other. They describe different stages of the same problem: one regulator collecting a cheque for a period that ended in 2020, another admitting that a ban introduced last December has not stopped a substantial proportion of under-16s from staying online.

What to watch

Three things. First, the ICO's resumed investigation into TikTok's algorithm and teenagers' data, which the company has stopped obstructing. Second, the UK's under-16 regime, due next year, and whether Ofcom gets enforcement powers that bite. Third, whether Sanders's proposed pause on advanced AI development moves at all, or joins the pile of similar bills that never reached a vote.

Gates's line about overhead is the one to keep. Regulation of platforms and AI is, for the companies involved, a cost of doing business. The argument is about who sets the price.