We might not understand artificial intelligence

We cannot recognize advanced intelligence.

 

This used to be a theoretical problem, but may soon become very practical. Because we started to build (artificial) intelligence.

Why?

Our only practical definition of intelligence is by similarity: the Turing definition: "it's intelligent if it's like us". And contemporary AI already passed the Turing test.

But consider this: lower levels of intelligence cannot SEE our intelligence. Plants, and animals, have a certain degree of intelligence themselves. 

We share a huge amount of DNA with animals and plants. And we have the same basic type of evolutionary, adaptive intelligence. Animals and even plants communicate, collaborate, learn, adapt, evolve, control their environment.

But plants and animals, although very similar to humans, are incapable to understand or even recognize the level of intelligence of humans.

They don't even know that they don't know that there is human intelligence.

An insect, or a fish, cannot recognize the manifestations of human intelligence: writing, roads, satellites, radio waves, wifi signal, music, abstract painting. This was superbly exemplified in Strugatki's novel Roadside Picnic. Therefore, we might be surrounded by an infinity of manifestations of higher/different intelligences, that we cannot recognize, or even detect.

So we cannot even recognize extraterrestrial, different, or superior intelligence.

And soon, we might not be able to recognize artificial intelligence.