top of page



I first studied Artificial Intelligence (AI) around 25 years ago. My AI project, at the time, was detecting defects in die cast components using some of the main AI tools available at the time. These included: rule-based systems which used ‘if-then’ constructs to make decisions, i.e., if this is true then take this course of action, e.g., if the door is open then keep walking; neural networks (computer algorithms) which were used for pattern recognition, e.g., if you look around your room you can probably identify most, if not all, objects in the room. This may be despite only seeing parts of some of the objects. The reason you know that a partly hidden table is in fact a table is because you have seen many tables in your life and you have stored these images in your memory. In identifying the table as such you are dealing with probabilities i.e., identifying the object partly seen as a table is due to this particular outcome scoring the highest probability score in terms of all other possible solutions. Another AI tool is the knowledge base. Here, knowledge about the environment is collected and stored in a database, e.g., the locations of objects in a room.

So, what makes something intelligent? In my view there are two main things: the ability, when acting alone, to learn and the ability, also when acting alone, to make decisions. In fact, much of what we consider AI is simply copying human behaviour. So how does a machine learn? Machines learn by using what they discover about their environment, combined with information they are given by the manufacturer or user, to build models or databases which then, in some way, describes that environment and enables them to interact with the environment in a useful way. Key to this, is the need to store data and information; and a lot of it! This means that huge amounts of storage capacity are needed (I am not writing here about machines which can only do simple tasks i.e., ‘dumb-machines’) as well as quickness of response. Speed of response brings in another AI problem; that of ‘search’. Search refers to the quest to find the best solution (see the ‘Travelling Salesman’ problem for an example of how difficult this can be) or at the very least an acceptable solution in a given time. The reality is humans do this all the time by gathering information using their senses and then, by a mixture of experience and reasoning, come up with a solution. Sometimes this is the best solution, but based on the complexity of the problem, often as not, it is simply an acceptable one. This is how we humans live and our ability to find the best solution to a problem e.g., should I take the number 27 or the number 35 bus home, is often what makes the difference between a good day and a bad day.

So, AI is here to stay and as machines become smarter, the need for greater storage capacity and faster response time will continue on an upwards trajectory. One solution to this is the quantum computer. Right now, a useable quantum computer is some way off, mainly due to its lack of stability and the fact that, at the moment, they only seem to be suitable for the few applications that standard computers don’t do very well; in other words, the computers that we use at the moment, including the one I am using to type up this article, perform the vast majority of our everyday tasks very well and will continue to do so for the foreseeable future.

The main difference between a quantum computer and a standard computer is to do with their possible data values and how data is physically stored. In a standard computer basic data unit, ‘1’s and ‘0’s, are represented by a high voltage and a low voltage respectively whereas in a quantum computer the same data may be represented by the spin of an electron (‘up’ or ‘down’) or the polarization of a photon (’vertical’ or ‘horizontal’). In a standard computer the most basic piece of data, known as a bit, can be either ‘1’ or ‘0’. In a quantum computer, a qubit, can also be ‘1’ or ‘0’ but it can also be ‘1’ and ‘0’ at the same time. This phenomenon, where something can have two different values at the same time, is called quantum superposition. Another quantum phenomenon is entanglement. Entanglement is when a pair or group of particles interact in a way that each particle is unable to act independently, i.e., a transformation of one particle will be felt by all other particles in the group. This is true even when the particles, which make up the group, are in different locations and this is where the phenomena may become useful in that data may be able to be transformed (modified) en masse by simply applying a transformation to a single particle. One area of research, and where quantum computers may prove most useful is cryptography, and it is this application, most of all, that may yet drive the development of quantum computing to one day provide the necessary computing power needed to control our next generation of intelligent machines.

Finally, and in a week when the existence of a possible fifth force of nature has been suggested,  (alongside gravity, electromagnetism, the strong nuclear force and the weak nuclear force), I believe it will be as a result of new discoveries in quantum mechanics that we will find solutions to our future technology needs; micro-electronic components which provides solutions to today’s technology needs, exist only because of the discovery and then understanding of how sub-atomic particles, such as electrons, behave. In considering the microscopic rather than the macroscopic world to look for solutions, it is simply the case that to see what is really happening that we have to look closely at the smallest possible level (analogy: to see if it is windy outside look at a blade of grass rather than a large building). Another issue which may be preventing progress is the obsession we humans have with building everything in our image or else re-vamping what went before. AI seems to be based on how we learn and make decisions and quantum computers seem intent on using quantum versions of long-established logic gates in order to make decisions. This may be the way to go but sometimes, with something new, the approach has to be a different one and it is in the world of quantum mechanics that such new methods may show themselves. And then the next real jump forward can begin.

AI machine (2).jpg


AI and Quantum Computing: Arts Articles
bottom of page