Are artificial systems genuinely intelligent?

There is a lot of scaremongering around AI. Most of this comes from companies trying to sell AI and therefore using inappropriate language designed to pull the user in. Terms such as ‘learning’ and ‘intelligence’ are not ones that we can apply to machinery. Heck, we don’t even understand our own brain or our own intelligence yet.

IST AI logo

Statistics, over the years, along with maths and physics, has been amalgamated into algorithms. Algorithms are, very basically, long programs combining many rules. The specifics that one has to understand in order to get a program to do even the most basic of things is outstanding.

When we program we put our own subjectivity into it, we don’t fully break down all of our thoughts or rules, because we never had to before and we don’t quite know how to now. If we tried to find the fundamental root of our beliefs or knowledge, we normally fail, because we just ‘know’. This ‘just know’ is based on a lot of information, evolution and just basic instincts. This is not something that can just be taken and put into a computer program.

What the ‘AI’ people are trying to say they are doing is construct humanity from metal. Why? This is Jurassic park-esque. I doubt anyone knows why. That’s because what they are really doing is using algorithms to automate basic tasks for us and/or optimise them in a bounded solution space. These programs cannot think or learn. They can only optimise. Just for a second, imagine what makes you, you. Imagine what makes you able to think and be self aware. Even eminent philosophers sometimes struggle to understand or express this because it is so complicated and we can’t step outside ourselves to answer in any reasonable way.

Your Hive ‘learns about your daily movements to make sure you are always warm’ – the language here is completely wrong. Your Hive programs a schedule of heating after having collected the data of when you are at home. Automatically we can see the difference between the languages. One makes us more predisposed to the thermostat, and one renders it simply a program. This anthropomorphism of a metal object is not only irresponsible, but it also works! It’s so easy to do and a method employed in sales to make you care about something.

But why do we want to anthropomorphise things? Well, we do it all the time, we see behavior like ours in animals, in objects, in abstract concepts, because our frame of reference is us and also, we are quite poor at being objective, we like things in terms we can understand.

The next time you see advertising question what’s going on and how you are being sold on something. A marketing video on LinkedIn that went viral showed a picture of pigs living in awful surroundings. The picture was recent and nothing had changed – the advertising campaign said something like ‘we are making advances in how we home pigs to make them happier’. The same campaign showed pigs being injected with drugs, the campaign said ‘making meat safer for human consumption and caring for the pigs’. All of a sudden the spin has made you think in a completely different way.

So before you run for the hills because ‘AI’ is coming to get you – consider the goals of the promoters of this.