Artificial Intelligence and Crime Detection

Questions from "The Question":

Artificial intelligence and detection of crimes by a group of people by prior conspiracy: when will this be possible and were there cases?

There are such technologies. For example, in 2017 it became known that VALCRI (Visual Analytics for sense-making in Criminal Intelligence analysis) is being introduced and tested in Britain. The automated computer system to combat crime, uses the latest technology of machine learning to analyze numerous police records, interrogations, photos from crime scenes, videos, and so on. Interestingly, this program follows in detail the deduction method described in Conan Doyle's novels about Sherlock Holmes.

Other examples are also easy to find, as neural networks are a good tool for many analytical systems, including law enforcement agencies.

However, there is a nuance. There is no clear evidence that Artificial Intelligence has solved the crime itself. And this is understandable. Expecting from such a catchy name the likeness of a "smart machine" we forget that in fact neural networks are, first of all, a new tool for humans. It is not some "smart independent machine". Thus, even with the use of AI in law enforcement agencies and special services, the final solution lies on the operator's shoulders. At least for now, those technologies that are successfully developed in the AI field work in this way, helping to automate the process.

When it becomes possible, it is impossible to say that it is not just to automate the process of work of police and other similar structures, but to make it less autonomous. By the way, not only in this field, but also in any other area related to AI, and that is why.

Experts on neural networks directly state that knowledge about the work of the human brain and consciousness is extremely insignificant. In fact, no one fully understands how our consciousness works. What does this have to do with AI? Everything is simple - creating a synthetic mind we will proceed from known knowledge and analogies, and therefore from the work of our own mind. And here, from the point of view of science, a lot is not studied at all, it is not clear how it works and so on. In other words, creating a truly synthetic mind based on AI at the moment we are like a child who is trying to build a shopping mall without knowledge and instructions. Of course we can make 4 walls and a nice roof, but for tens of thousands of people such a "handicraft" will not be suitable.

So all "analytics" and forecasts on this topic are nothing more than speculation. In order to start to really predict the emergence of such technologies (self-sustaining AI), you need to cross the boundary of applied development. To make the first sample of an independent and intelligent machine. And not just within the framework of a laboratory - but applicable successfully in the real life of people. Alas, there is a lot of talk on this subject, but in reality - all such developments are at the stage of experimentation (and sometimes even theories). Everything else that we see in our smartphones, computers and other areas marked "AI" is actually applied tools based on neural networks, which will be useless without the operator (but certainly already now significantly simplifies human life by successfully automating processes).

Author: Sergey Ulyashenkov
Technologies >