defining "intelligence"

The below highlighted paragraph is my response to a comment on an interview of me by Taz, that lectured me I didn't know the definition of intelligence since my definition (doing the right thing at the right time in response to a dynamic environment) doesn't include the clearly implied ability to learn.
Even outside of technology there is a debate about these kinds of terms. I think whenever a machine (or organism) converts perception to action it is worthy of notice, so I call this "intelligence". When an intelligence can modify itself (that is, learn) I call it "cognitive". But this is bending folk-concepts to a technical use for the purpose of communicating fine points.  Of course you can find people who when they say "intelligent" mean "conscious" or "human" or "moral" or many other things.  Science though is about taking concepts apart so we can be precise.
I learned as an undergraduate that where definitions are not generally agreed, what you need to do is to lead off a paper or discussion with the definition you are going to use, just for clarity, not to say that's the "right" one. In general, I favour reductive definitions like the above because they are easier to both explain and defend, but that doesn't mean that an academic can assert what you as an individual means when you say a word.

With respect to the actual content of the Taz interview, here is an old (2013) blog post where I go into more detail about why I believe we've had superintelligence for 10,000 years, implying the intelligence explosion was ages ago.

Green plants as intelligent organisms is by Anthony Trewavas, photo is by me.

My favourite tweet about the Taz interview is:  Jo, Intelligenz ist gefährlich, aber diese Singularität ist vor 10k Jahren passiert.  :-)

If you want to cite my definition, it appears in a lot of my writing, but most recently in the article Standardizing Ethical Design for Artificial Intelligence and Autonomous Systems.  Green open access version at Bath.


0