Robert Leach examines whether artificial intelligence and machine learning are simply ‘tech jargon’, or valuable decision support tools that we need to better understand.
Technology is continuing its relentless progress, promising to make everything we do easier, quicker and less costly, making us more productive and enabling change. Phrases and acronyms abound, and technologists and commentators alike use them all the time, assuming – perhaps wrongly – that their would-be customers really understand what they mean: artificial intelligence (AI), machine learning (ML), Internet of Things, etc.
While the techies may understand the application of this new technology, the rest of us need to know – in our terms – what it can do for us; how it can help resolve our day-to-day challenges, address our future plans or help us practically, by increasing the capacity and capability of our teams, driving faster results and better outcomes.
Consider, for a moment, that you have a 60-page Word document to read. You look at the index and see a chapter of interest, but you do not want to read the whole chapter, never mind the whole document, so you click on the ‘Find’ button and enter key words or a phrase that should lead you directly to where you want to be.
What you have done is apply a form of AI to help you be more productive, to arrive at the point in the document that is most relevant to you, enabling you to consider information and take an appropriate action.
For example, our investigative team is trying to establish the historical movements of a suspect as they walk from one place to another. You do not know anything other than the date, start location and an estimated time, and the end location. What you would rather not do is log in (if you could practically do this) to the various CCTV cameras in the vicinity of all of the potential routes the suspect could take, and then employ a team to monitor each set of cameras to see which route the suspect might have taken. Ideally, you would like to get straight to the point where you are sure, beyond doubt, that you know the route the suspect took.
A combination of camera outputs and software technologies working in the background – let us call it ‘recognition technology’ – can do the hard work, the ‘heavy-lifting’ involved in processing masses of information quickly. Technology can be ‘trained’ to do huge number of things virtually instantaneously. What the technology gives you is optional views of what went on based on the content from the cameras across each route – it is then that your expert investigators decide what to do with the information.
If you knew certain features about the person whose movements you are interested in, different software could help identify that person (or someone similar), to a reasonable level of confidence, so you could home-in on the route they took. What the technology cannot do is decide whether the person or people you are seeing is (are) your suspect(s) – you make that decision. All the technology does is give you what it sees based on what you tell it – albeit very quickly. It gives you the options, you draw the conclusions.
Most technology we use is effectively a ‘decision support tool’ – from key word search to something more complex. For example, modern case management systems draw the user through the correct parts of the complex processes in building towards a conclusion. They are being led by the software. The latest software ‘learns’ what parts of the system you are using the most and steers you towards these.
Other software solutions, such as Hitachi Digital Evidence Management, reduce the complexity involved in assimilating masses of data and present the often vast number of evidential exhibits in one place (in an ‘evidence basket’) to enable analytics techniques to be used to achieve huge efficiencies in the investigative process – including giving the Crown Prosecution Service access to the evidential strands of the investigation to enable better outcomes, including minimising the disclosure risk.
Hitachi has also developed sophisticated camera and software technologies that can identify and find people without any need to use facial recognition. These rather clever techniques can find a person whether they walk or run, whether they carry a bag or not, even on consecutive days when they have changed what they wear – and not requiring to view, or match against, any facial images.
Still more technology can run countless searches – defined by you – on this torrent of data to allow you to make whatever operational or business decision is required based on your confidence in the resulting information.
These are the types of things AI and ML can do for you. Decision support tools may be getting more sophisticated and cleverer, but they still do not make the final decision. They give you information, they give you options and they can provide ‘confidence levels’ – quickly. But you decide what to do next.
Decision support tools provide informed data to support decision-making. Their use can therefore increase the capacity and capability (and productivity) of your existing resources, for example, to carry out many more investigations per team than is possible today – and with the crime stats and continuing pressure on resources, this should be a very good thing to encourage.
There should be no concern that the ‘tech’ will reduce either the need for, or expertise of, policing’s resources – the experts still have to make the decision based on what the technology is suggesting. In fact, given that the technology never sleeps and capability will continue to improve, which will lead to ever more – and ever more reliable – output, we may actually need more specialists to respond to what the technology is suggesting.
As next week’s Excellence in Policing conference programme illustrates, in striving to balance escalating demand for services with diminished resources, police are making difficult choices – and in making choices there is a daunting number of things to consider.
Proper use of all this new technology will undoubtedly help.
Hitachi is a main sponsor of the Excellence in Policing Conference 2018, being held at the College of Policing Conference Centre, Ryton on Dunsmore, on September 25 and 26.
Robert Leach is the former chief executive officer of the Police ICT Company and now Head of Police and Public Sector at Hitachi.