Artificial General Intelligence has long been the dream of scientists for as long as Artificial Intelligence (AI) has been around, which is a long time.
Artificial Intelligence has had its ups and downs. There was even what many observers called an AI Winter, when investors decided to look elsewhere for more exciting technologies.
In recent years we have been getting quite good at Artificial Intelligence. We now have the mechanisms and techniques in place to adequately ‘train’ an AI to do a linear job pretty well. And from mortgage advice to customer service and HR, AI is being deployed in many different environments.
Yet we are still a long, long way away from anything that resembles Artificial General Intelligence, where AI mimics a brain, capable of reasoning and performing multiple tasks while understanding the context.
This, of course, may be a good thing. If we had an Artificial General Intelligence that worked now, we might find we had real problems. Already, with ‘linear’ AI, we hear an almost constant stream of stories of bias. Working on the old-fashioned theory of all things computing, you put rubbish in, you get rubbish out. In the same way, many AI systems are skewed towards certain biases, such as race or gender.
Imagine an Artificial General Intelligence system that was biased. And even if we found ways of ‘training’ such a system to be unbiased (which would be almost impossible) if such a system was confronted with the world as it is, a very different approach would be applied to problems such as global warming, various conflicts and even the current pandemic.
Logic or an Artificial General Intelligence system might come to the conclusion that global warming was the most urgent agenda item and that humans are the cause. Therefore, humans are harming their environment, therefore humans should go.
For now, Artificial General Intelligence remains the dream of a few scientists who want to go ‘all the way’ and one day, we may have the tools to build such a thing.
Let us hope we are ready (and are treating our world a bit better). In the meantime, let us continue to develop better AI, that can do simple things really well. Like, write articles about itself.