The Fierce Debate on General Artificial Intelligence

Image by Gerd Altmann from Pixabay

In the world of Artificial Intelligence (AI), there is a raging debate on whether or not we can achieve General Artificial Intelligence (GAI). Some believe that it will happen in the next few years, others that it will never come about. Those opposed have some fairly solid arguments as to why not.

So what is GAI? Essentially, it is when AI becomes “sentient”, can truly think on its own, formulate ideas, views and opinions and be an equal to mankind in intellectual terms. Perhaps even smarter. Right now, AI, an umbrella term for many different technologies, is known as Narrow AI (NAI.) AI is very good and getting better all the time, but it is really only good at doing one thing. There is no sentience, no consciousness.

And while much progress on AI has been made, especially in the last few years, much of it remains more hype than reality. AI is an extremely useful technology and will likely play a vital role in improving healthcare and helping us fight complex issues such as climate change. Humanity may only be able to fight climate change with Artificial Intelligence.

But as AI gets ever better in doing certain jobs, we are learning some unintended consequences. Two of the biggest being how racial and gender bias are inherently coded into AI. Not with intent, but it comes out because of existing, systemic and cultural biases within human society. This has lead to a growing field of AI ethics. A field that is causing no small degree of friction between social justice advocates, government and industry. This in large part, is sparking the debate over General Artificial Intelligence (GAI.) Do we want to go that far and can we even achieve it?

To get to GAI some rather large hurdles have to be overcome. And when I say large, I mean problems we just haven’t been able to solve yet. Maybe someday we will. Perhaps Narrow AI will help solve them? They are;

• Consciousness: We do not know what this is. Why and how are we conscious? What is it exactly? There is a lot of work being done in this area for good reason. Some researchers, in this paper, think it is an entirely logical process and are building a framework. But some things may not be code-able. Then there’s quantum physics using the study of brain changes under anesthesia. For now, consciousness remains speculative at best.
• Intelligence: We haven’t really defined what intelligence actually is either. Ironic. Oddly enough as well, it doesn’t get as much attention as research into consciousness. Even though that should, perhaps, come first? Some new research suggests a gene mutation causes higher intelligence. It’s very complicated.
• Irony and abstract thinking: A significant aspect of consciousness is our ability to understand abstract and indirect meanings such as irony. Machines cannot do this. Nor can machines understand emotions or understand a human’s mental state. Essentially, we understand each other through analogy. Computer scientists can’t program this. Mostly because we still don’t quite understand how we understand each other reliably.
• Human: Even if we do discover what consciousness and intelligence are and we get computers to “think” (we don’t really know what that means either), they aren’t human. Logically, then, AI can’t think like a human because it is not.

We may or may not get to GAI at some point. But until we can solve for the problems above, it is unlikely. We can build some amazing AI technologies that will truly help us build a better world. We are also weaponizing AI and in a world increasingly divided, GAI may not be a very good idea. But humans being humans, we will continue to work towards GAI. That pursuit in itself may not produce GAI, but will certainly help AI to progress in some impressive ways.

To get to GAI, we have to unravel some very complex and deep mysteries. We must solve some very philosophical ideas that we’ve been contemplating for centuries. We cannot code for them because we don’t know what to code for. It might help if computer scientists and engineers involved the social sciences more, but this is rarely done in the field of AI and when it is, it is mostly at an academic level. But that’s another topic entirely.