These four assumptions are almost always made in science fiction and popular commentary on AI or alien life. They are often made by the leaders of AI research and industry, as well as by academic futurists and professional scientists (who should know better).
- There is some essential aspect of a human being that gives rise to all the important aspects of a human being.
In science fiction, the essential property is called things like “intelligence”, “consciousness”, “sentience”, or “self-awareness”.
- All things in the universe have an essential aspect that has a linear measure.
The measure is described metaphorically as “altitude”; there are “higher” and “lower” things. It generally respects this ordering: inorganic objects < microorganisms, fungi, etc. < plants < non-mammal animals < non-human mammals < humans.
- Technological capability and human knowledge increase at a (roughly) consistent rate.
The Myth of Limitless Progress
- Technology will eventually overcome any limit. Any form of magic is possible with sufficiently advanced technology.
Hubert Dreyfus once said, about similar assumptions, that it is “taken by workers in AI as an axiom, guaranteeing results, whereas it is, in fact, one hypothesis among others, to be tested by the success of such work.”
I will show that we intuitively accept these assumptions when we’re watching movies, or, sometimes, when we speak before we’ve really thought things through. I want to establish the existence of these tropes, the reality of their presence in our culture. I’m primarily doing a kind of anthropology — a study of how we think about AI.
Sometimes these assumptions don’t square with obvious facts. Sometimes there are contradictions and paradoxes in the way we understand them. In either case, they can’t be true in the way we normally understand them. I’m not capable of resolving these paradoxes and that’s not my main purpose (although sometimes I can’t resist trying). But I can help frame the questions and dismiss the most naive answers.
CHARLES GILLINGHAM 