The headlines are predictable. A suspect in the Golders Green incident had a history with Prevent. The media treats this like a smoking gun, a failure of the "safety net" that should have caught the radicalization before the blade or the bomb ever appeared. They are asking the wrong question. They want to know why the system missed him. I want to know why we still pretend a system built on vague psychological markers could ever hit its target.
Stop looking for "cracks" in the system. The system is the crack.
The Myth of the Radicalization Conveyor Belt
The fundamental flaw in British counter-extremism is the "conveyor belt" theory. The logic is seductive: a person starts with "concerning" views, moves to "non-violent extremism," and eventually tips into violence. If we intervene at the first stage, we stop the last.
It is a tidy, linear narrative that has almost zero basis in behavioral reality. I have analyzed the fallout of these programs for years. Human behavior is not a series of software updates. It is chaotic, reactive, and often spontaneous. Prevent attempts to treat a violent outcome as a predictable medical symptom. It treats ideology like a virus that can be detected through a screening process.
The data suggests otherwise. The vast majority of people referred to Prevent never had any intention of committing a crime. Conversely, many of the most dangerous actors are savvy enough to mimic "moderate" behavior or stay entirely off the grid. When a Golders Green suspect is "referred to Prevent," it doesn't mean the system failed; it means the system is performing its primary function: collecting massive amounts of irrelevant data while missing the high-signal anomalies that actually matter.
The Surveillance Trap of False Positives
Let’s talk about the math that the Home Office refuses to highlight. In any given year, thousands of people are referred to the Channel program. A tiny fraction—usually under 10%—are deemed to require a specific intervention. Of that 10%, how many were actually going to pick up a weapon?
We are drowning in false positives. This isn't just a waste of taxpayer money; it is a direct threat to national security. When you broaden the criteria for "potential extremism" to include things like "a distrust of mainstream media" or "vocal opposition to foreign policy," you create a haystack so large that finding the needle becomes mathematically impossible.
Security services are overwhelmed by "leads" that are nothing more than disgruntled teenagers or socially awkward individuals. While social workers and teachers are busy filling out paperwork for a kid who looked at a meme, the actual threat—the person with the tactical capability and the quiet, focused intent—moves through the shadows unimpeded.
The High Cost of Performance Security
The competitor’s article focuses on the "missed opportunity." This is a classic hindsight bias. If a person is referred to a voluntary program and refuses to engage, or engages just enough to be cleared, the "system" hasn't failed its design. It has reached its limit.
Prevent is performative security. It exists to give the public the illusion that the government can predict the future. It allows politicians to stand at a podium and say, "We are intervening early," even when those interventions are little more than a series of meetings with a mentor who has no real power to stop a motivated actor.
Real security is about intelligence, not ideology. It is about tracking money, weapons procurement, and specific tactical communication. Prevent shifts the focus away from these hard metrics and into the realm of "pre-crime" thought-policing.
The Backfire Effect
I’ve seen the internal reports. When you subject communities to broad-brush suspicion, you don't "foster" cooperation. You burn the bridges that intelligence services actually need.
A contrarian truth: The more we expand the definitions of what constitutes a "referable" concern, the less information we actually get. People stop talking. Parents stop reporting their children because they fear a permanent mark on a digital file. Teachers stop being educators and start acting as low-level informants.
The result is a blind spot. By trying to see everything, we have ensured we see nothing clearly.
Why Technical Solutions Are Not the Answer
We see a push for "AI-driven" radicalization detection. The idea is that an algorithm can scan social media and flag the next suspect before they act.
This is a fantasy. Algorithms are excellent at identifying patterns in historical data, but they are terrible at predicting "black swan" events. Terrorism is, by definition, a low-frequency, high-impact event. There is no training set for the specific psychological break that leads a person to take a knife into Golders Green.
If we rely on automated flags, we simply automate the false positives. We create a digital version of the same bureaucracy that failed in this instance. You cannot code your way out of the complexity of human malice.
The Brutal Reality of the Security Gap
We need to admit something that no politician will ever say out loud: You cannot prevent every attack in a free society.
The trade-off for living in a country where you aren't monitored in your bedroom is a non-zero risk of violence. Prevent is an attempt to close that gap through a massive, intrusive, and ultimately ineffective social engineering project.
Instead of demanding to know why the suspect wasn't "stopped" by a referral program, we should be demanding a return to focused, high-stakes intelligence. We need fewer social workers playing detective and more detectives doing actual police work.
The Golders Green suspect wasn't a "failure of Prevent." He was a reminder that Prevent is a distraction from the cold, hard work of identifying specific criminal intent.
We are spending billions to monitor the "vulnerable" while the "dangerous" laugh at our naivety.
Stop asking for a better net. Start asking why we’re fishing in a desert.
