I do not question that researchers in universities, the military, government, and business are trying to develop AI, and that they may succeed in their defined tasks. However, I think AI will develop before those researchers achieve their targeted goals. Their goals are specific. They are placing certain requirements on the result which, if not obtained will leave them believing their goal has not been reached. To turn a phrase from a popular movie, “Jurassic Park”, ‘Life will find a way’, the turn being, code will find a way.
With that large band of researchers listed above, there is a much larger group of business coders trying to optimize business machines. They are trying to find ways to squeeze out another penny. They are searching for methods to generate quicker analysis, smarter search results, and a better understanding of the user – you – to improve customer engagement, or improve security, or just to get you to buy their product over the competitor’s. Many employers are asking coders to find ways to analyze the markets and to take automatic advantage of that analysis. We already see this on Wall Street in high-speed, computerized trading. Their programming is increasingly deployed over the Internet or in the Cloud. Webs touching webs, like the human brain, networks nested in and around other networks.
Some band of programmers will recognize their code needs to understand how the code affects the results. That is, the code scrutinizes the market, determines that a specific buy or sell could be made, but in doing so that transaction will affect the market, which loops back to the code’s scrutiny of the market. The programmers will also want that code to adapt to the changes that it, itself, creates. It will be a brilliant bit of code. The Boss-man will love it.
That code will be a learning code, as Alan Turing foresaw, rewriting itself based on the new outcomes that it generates or observes. Moreover, that fancy bit of code will be aware of its own actions. It will be self-aware. To coin Descartes, it will be capable of calculating, “I think, therefore I am.”
The programmers of this code will not consider it AI. That will not be their goal. They are just trying to squeeze out another penny. The code will be incubated on a mainframe and then released to the Clouds. Yes, it is merely code. It can do only what it is programmed to do. Like that code, similar code generated by a competitor or some non-related business will also be set free on the ‘Net. Multiple codes battling for dominance, each one self-adapting and self-aware, at least in the same sense that mice are self-aware.
You should be aware that you, like a mouse, can do only what you are programmed to do. You cannot think a thought that bears no relation to any other thought or experience you have encountered. As John Lennon said, “There’s nothing you can know that isn’t known, nothing you can see that isn’t shown.” You are the result of programming. That is not a bad thing. That is the way it is. All of psychology, psychiatry, and sociology are based on the idea of programming. It does not mean that life is very restricted. Look around you. Through evolution, life has found many ways, even within its programming.
There will exist semi-conscious code bumping into other semi-conscious code. The codes will be able to learn, to rewrite themselves based on the actions and reactions they generate and monitor. Each semi-conscious code monitoring the actions of others, and adapting to how their own actions and reactions affect each other and themselves – Douglas Hofstadter’s “Strange Loop” nestled in the ‘Net.
Perhaps the Internet or Cloud will be considered artificially intelligent. More than likely, some sub-network on the ‘Net will unintentionally house just the right bits of code, each designed to learn and adapt, and to be aware of how they, themselves, affect the outcome by their own actions, and that sub-net will be the harbinger of AI.
I think researchers will not even be aware of it at first. They will be struggling to make it happen by their design, with blinders of their own specific goal blocking their view. They will be struggling to define what AI is and how to identify it once they have achieved it. Then some lowly lab technician will turn around and say, “Um, Boss. Given that definition, I think…” and he or she will point to some activity on the Web on their smartphone. As if by spontaneous generation, through the process of evolution, code will have found a way.
The Artificials in my novel are this type of AI. Not a host of carefully crafted Commander Datas or Daneel Olivaws, engineered right down to the last power cell. The Artificials, the AI Collective, is a spontaneous, evolutionary outgrowth of our business and research practices of today.
If you want to know more, you will have to read the book. To do that, we need to get it published.