Looking first at its history, more than 25 years of study into the impact of CCTV on crime generated an extensive body of evidence but the jury has yet to reach unanimity on a verdict. However, all of the research was undertaken when cameras just filmed things. Most modern surveillance devices are not cameras anymore; they are networked computers, one basic function of which is to capture images. They are rarely on closed circuits and they are not on TV.

The received wisdom on the impact of public space surveillance took place before the camera knew who it was looking for; before it 'knew' anything in fact. Now the AI-driven device looking for me knows exactly what I look like and also what I sound like, how I walk, what I wear, drive or ride, and not just at the 'time in question' but generally. It knows where I have been, where I go, when and with whom – again, not just on the night in question but habitually. It can communicate, it can teach and learn from other devices, it can retrieve, combine, synthesise and suggest and each time it looks for me it gets better at finding everyone. And I – the person being sought – know this too. This is nothing like CCTV.

As to its future, we know from retailers the significant difference that facial recognition systems has made in protecting their staff from assaults, abuse and harassment, reducing incidents dramatically (sometimes >60%) when compared to conventional in-store CCTV; it is a reasonable hypothesis that such intelligent devices being used by the police will have some direct positive impact on wider crime and criminality in way that was simply not possible or imaginable with CCTV.

Technology has also changed surveillance behaviour, both that of the police and the citizen. Look at the first thing police forces do when something important happens. The police response now – not just in the UK but in every country I have worked with – is to put out a call for people to send in images, anything from their GoPros, doorbells, dashcams, anycams. Why? Because someone almost certainly captured something relevant.