"MICHIO KAKU: What computers need is data. Data by which we can draw conclusions. And when we talk about intelligence in outer space, we have to be open to the idea that their intelligence may be quite different, quite different from our intelligence. And therefore, the question is how do we tell? how do we tell where real intelligence is, given the fact that there could be more than one type of intelligence?"
"- For the Science Channel, they once put me in a swimming pool with dolphins. Sensors picked up the squeals and chirps from the dolphins as I was swimming next to the dolphins, and then they ran it through a computer. The computer looked for regularities in the chirps and squeals of dolphins, realizing that they speak a totally different language than humans. And their criteria for intelligence would be different from our criteria of intelligence. So for example, the letter E is the most common letter in the English alphabet. And you can rank the frequency with which certain letters are used to pinpoint who wrote the article."
Current definitions of intelligence are biased toward human language and behavior and may miss nonhuman or extraterrestrial intelligences. Computational identification of intelligence requires large, relevant datasets and appropriate pattern-detection criteria. Animal examples, such as dolphin vocalizations, show distinct regularities that standard human-centric metrics would overlook. Multiple types of intelligence could exist, each using different modalities, signaling strategies, or information structures. Effective search strategies should broaden signal spaces, consider nonlinguistic regularities, and develop novel metrics capable of recognizing intelligence expressed through unfamiliar channels or architectures.
Read at Big Think
Unable to calculate read time
Collection
[
|
...
]