An extraordinary move in Manila this morning: the Philippine government has formally banned a popular video game following its discovery on the phone of a teenager who opened fire at a high school in the capital, killing four. This is a stark, real-world collision of digital cultures and public safety. While details remain scarce, the game in question is reportedly a hyper-violent battle royale title favoured by millions across Southeast Asia. President Ferdinand Marcos Jr. stated the ban is a temporary measure to allow for a 'thorough investigation into the gaming ecosystem's role in radicalising youth'.
Meanwhile, in London, Whitehall is quietly dealing with a different but equally pressing digital menace. UK security agencies have begun a formal assessment of risks posed by advanced AI systems. Sources say MI5 and GCHQ are running red-team exercises on large language models, testing for vulnerabilities that could be exploited to generate disinformation on an industrial scale or even to design biological weapons. This isn't science fiction. This is the blunt, administrative reality of 2027.
There is an invisible thread connecting Manila and London here. Both governments are waking up to the fact that software is now infrastructure. A video game is no longer just a toy; it is a vector for ideology, a training ground for violence. An AI isn't just a clever chatbot; it is a potential weapon of mass persuasion. The Philippines' ban is blunt, a sledgehammer approach. But the UK's method is scalpel work, discreet audits of code that could shape our elections and our psyches.
Consider the human cost. In Manila, four families are burying their children today. Did that video game cause the shooting? It's a complex, uncomfortable question. Correlation is not causation. Millions play the same game without ever picking up a real weapon. But the data is mounting. The algorithms that drive these games are designed to maximise engagement, often through aggressive, reward-based loops. Could that same loop, when combined with isolated depression and access to a firearm, become a kill switch? We don't know. But we cannot afford to wait for certainty.
In London, the calculus is different. The risk with AI is not a single event but a slow, corrosive drip. Deepfakes that erode trust, automated systems that deny people benefits, chatbots that radicalise lonely individuals in private Telegram groups. The UK's assessment is just that so far an assessment. But it signals a shift from treating AI as a technological marvel to treating it as a regulatory object, something that must be stress-tested like a bridge or a vaccine.
What unites these stories is a crisis of governance. Our laws were written for a world of atoms, not bits. How do you ban a video game that originates in a different jurisdiction? How do you audit an AI whose code is a proprietary secret? These are not problems with simple solutions. But the first step is admitting that the status quo is dangerous.
For the tech industry, this is the moment of reckoning. The era of 'move fast and break things' is over. Now, every product launch must consider the potential for real-world harm. The Philippines is saying 'no' to a specific game. The UK is asking 'what if?' about a broad category. The rest of the world is watching, and they are taking notes.
As we go to press, a young man in Manila has confessed, but his motivations remain a mystery. In London, a committee of civil servants is reading through a report on model weights and safety thresholds. The future is being written in two very different languages: Marcos's executive order and GCHQ's vulnerability log. We would do well to read both.
