Artificial intelligence (AI) could be a big issue in extraterrestrial First Contact. The first aliens we meet may be mechanical creations controlled by artificial intelligence. Thus far on this blog I have considered primarily the alien side of artificial intelligence: but what of human AI developments? Tad Friend has an excellent piece in the New Yorker that explores worries over the rise of human designed AI.
Researchers use three primary terms to describe human AI:
-Artificial Narrow Intelligence (ANI) is the term used to describe technology such as Apple’s Siri and self-driving cars. We have this in varying degrees of intricacy now.
-Artificial General Intelligence (AGI) is defined as being at the level of human intelligence. It is the goal of many companies and institutions currently working in the field. It is generally agreed that AGI has not yet been accomplished, although the boundaries are being pushed daily.
-The ultimate level of AI is called Artificial Superintelliegnce (ASI). It would be intelligence beyond any human capabilities, even for the most gifted of humans.
Defining AI based on human intelligence could be tough in the future, especially if we find ways to boost our own intelligence artificially. One could argue that your smartphone is a first step for that intelligence boost. What happens when your smartphone is directly connected to your brain? Of course AI researchers would say that access to information is not intelligence. Intelligence in the narrow definition is the ability to solve problems. Some would argue that a true test is the ability to understand shadings of meaning in language and use imagination to solve problems. Our decision to base AI levels on a comparison to human intelligence is perhaps a symptom of our anthropocentric thinking. However, it is also the only measure we currently have. Other definitions are not human based, but rather scale based- weak AI versus strong AI. This of course, would create problems as technology progresses. What is strong AI one day could be weak a few years later.
There is further debate with the suggestion by pioneering AI researcher Judea Pearl that we not use reasoning by association to rate AI (simply looking for correlations in data) but rather causal reasoning; inquiring how causal relationships would be altered if there is intervention. Kevin Hartnett has a story explaining this in the Atlantic Monthly.
So, let’s get back to the aliens. I have said before that there could be two basic types of alien AI machines exploring the universe: Biologically Originated Intelligence (BOI) and Artificially Originated Intelligence (AOI). The difference is simple: did biological beings create the intelligent robots cruising through space or did other artificial intelligence create those mechanical explorers? The answer has big implications for humanity. We worry about the rise of AGI and ultimately ASI. Will humans become extinct? Will we morph into increasingly mechanical beings? Will ASI decide to get rid of us or perhaps leave us behind to explore the universe while we struggle here on Earth? Those questions are far-fetched considering our current level of technology. But the concern it would raise in First Contact could have a direct impact on our relationship with alien AI. We could well understand a sophisticated alien probe controlled by biological creatures. However, a probe with AGI or ASI capabilities would be a concern. There would be an inherent threat involved in any alien machine visiting our solar system. The worries would increase as a direct correlation with the differences exhibited by that visiting alien life form. A big question could be the relationship between the original alien biological creatures and their created AI. Do they exist together in harmony? Or did the AI grow to supplant the biological beings? If the later is the case - that would likely create a great deal of concern among humans. We could find ourselves with some major issues to consider, ranging from what kind of contact we would want to have with alien AI to how much further we want to go with the development of human created AI. One could imagine quite a bit of angst on the part of humans. Certainly it would be a tough way to start a relationship.
And perhaps that is the reason aliens have not said hello yet: they are in fact advanced AI and don’t know if we can handle the idea or the threat. Alien AI might be better off waiting until we are further along on the evolutionary scale, if that is indeed where it leads.
I think the best message to humans under such a First Contact scenario would be this: we don’t have to follow the historical path of aliens. We are early enough in our AI development to choose a different road, perhaps with more closely controlled AI. Before freaking out we should carefully study alien history. It could show that the biological creatures moved willingly, over time, to increasingly mechanical based bodies. Perhaps AI merely assisted the biological intelligence until the two became indistinguishable? We would certainly want to request a timeline of alien history as part of our initial contact.
In the meantime, we need to keep using that human imagination. We can create more fictional stories that explore these issues. It is our best way to conceptualize such matters. One can scoff at books and movies as mere entertainment, but when the idea behind such stories has weight, and the ideas are thoughtful, it may be our best way of considering how we want to proceed with AI.