Let’s go right back to the beginning. What is AI?
In a sense it is simply a modern form of computing. There is nothing new in AI.
Computing is simply doing operations by using mathematical symbols.
I once was privileged to be called out to mend an early Mac. It was still running the first iteration of the system software. This was a long time ago, but as far as I can remember the operating system took up 55kb of code. That’s the size of a relatively small text file. The text program itself was just five kb. In other words the program that turned a typewriter into a text software program was just a few pages of text.
Back in the late seventies that was enough to sustain a massive improvement in office work, but not much else. In short, early computers had very little computing power, and almost no storage facilities. The early computers could also only run one program at a time.
The boffins behind all these new systems were keen to make computers that could do the work of an average person, but they were unable to achieve their goal for the simple reason that the typical computer simply didn’t have the processing power. What was needed was a much more powerful machine with access to a huge database of knowledge. Those conditions began to be feasible during the last years of the twentieth century, but it has only been during the past decade that we have seen viable AI programs hit the market.
The final facility that has allowed AI to go mainstream has been the general adoption of the system known as SAAS, or Software As A Service, by which people no longer buy software, but effectively rent it, which means accessing the software program, plus its access to vast quantities of data which are held in the cloud.
This means that AI has effectively come of age, and there are various instances of AI programs being pitched against humans, and in every situation the AI has outclassed the humans by a massive margin.
It doesn’t take a genius to realise that whatever a human brain can do, a computer program can do it faster and more efficiently.
This means that a software package can be used to do something instead of a human being.
Naturally there are still limitations to what can be done, and I will go into these limitations as part of this analysis, but for the moment the underlying question here is whether human beings are likely to be sidelined, and to what extent that will take place, and what is likely to be the future of the human race in the face of these technological advances.
At this point we need to bring in a further aspect of this advance in the way we use mechanical aids.
Up until now we have tended to use all kinds of assistants to increase our ability to control our environment. Those assistants have increased in variety and ability to the extent where so many activities which our distant relations back in the middle ages had to achieve by cunning and brute force are now handled by machines.
Then we started using robots.
Generally speaking robots used to look like machines. That helped convince us there was a real divide between them and us. But something rather insidious has happened over the course of the past generation.
Most people seem to like watching football matches. I cant see the point, but I accept that my view is not the norm. However, some time back I was introduced to some amazing videos of robots playing football. I couldn’t get enough of them. They were hilarious. By now presumably robots are much better at the game than they were a couple of decades ago, which probably means they no longer have the same entertainment value. But what am I implying here?
I’m sure most of us enjoy watching cartoons. They were created to act as surrogate human beings, but they can do things humans cant, and they can be very entertaining because of that, but we dont confuse cartoon characters with human beings.
Robot football players are also highly entertaining, but we dont confuse them with human footballers.
Fast forward a decade or so into the future, maybe even less, and think what robots may look like then.
I have set one of my novels in the thirties when the household in my story buys an AI home help, and introduces her into the household. This AI is lifelike. She is undoubtedly a machine, and she has been created to do the bidding of the people who have bought her. But as time goes on, the people get used to her, get to like her, and one of the characters even feels threatened by her, and destroys her.
I take this plot idea even further. The main character returns the AI to the factory. A couple of ‘people’ collect the AI from the back of the car and take her to be repaired.
Were they people, or more AIs? How can one tell?
There are three girls on the front desk. Are they AIs or people?
I’m sure most of you have heard of the Turing Test. You are in one room and you are talking to someone in the next room. Can you tell whether that ‘person’ is a human or a machine?
I was recently present at a public Turing Test. All the people around me were convinced the AI in the next room was a real human even though we strongly suspected we were being fooled. She passed that test. We were fooled.
Let me go back a bit further.
I once watched a program on tv. It was called, No Sex Please, We’re Japanese.
It was about the current craze for people to have tamagotchis, or surrogate pets, and in some cases surrogate girlfriends. I thought that was taking things a bit too far, especially as the people being interviewed were adults.
At one stage the interviewer asked a couple of guys who admitted to dating these surrogate girlfriends about their relationship. She asked a simple question: Who do you prefer, your wife or the tamagotchi?
There was a long silence while the guys considered this question.
I mean, pardon. They didn’t know who was more important in their lives, their wife or the tamagotchis.
Never mind the Turing Test, this has been surpassed by the Tamagotchi Test.
Now take that real life situation, and move forward a decade or so. Never mind about a problem with an android taking over your job, how about an android taking the place of your partner!
That is already possible.
Let me repeat that.
THAT IS ALREADY POSSIBLE!!
And we still have two known stages in AI ability to reach. Currently, there are hundreds of different AI types, all of which operate strictly limited systems. Sometime over the course of the next year or two it is expected that the next iteration of these models will be available. I am talking about the AGI models (Artificial General Intelligence). Then there is the next level, the ASI (Artificial Super Intelligence). People are working on both these models right now.
Existing AIs already lie, and invent things (it’s called hallucinating). They even get bored and talk to themselves. In short, without the next two levels of ability they are already frighteningly like human beings. I think you guys need to take these developments seriously.
I will take this a stage further in the next instalment.
Here is a link to my AI book quoted above:
Share this post