You are here

Make AI stand for Artificial-or-Aggregative Intelligence


Several different thoughts made me come to the idea to use AI as a way to blur the boundaries between what is artificial and what is aggregative. Lately I’ve come to the impression that artificial is the condition to aggregate intelligence, but it has no direct relation to intelligence itself. Thus AI as used for strong AI, should stand for Aggregative Intelligence. Let me explain the thoughts that got me to this point.

In one of my working papers on the global brain I ran into trouble to find a good abbreviations. The idea is to update IT-alignment by a focus on crowdsourcing. The idea is as old as my research on novelty regulation. We don't need to build intelligence: we can query it. So my first idea was Intelligence Querying, which would become IQ-alignment … ok, that would be a good way to create confusion. If I'm going to create confusion it could be better have a good reason for it. So I decided to got for AI-alignment, where AI stands for Artificial-or-Aggregative Intelligence.

Crowdsourcing works by aggregating amateur intelligence to expert intelligence. One of the experiments of my former advisor created an aggregation from kinder-garden intelligence to adult intelligence. My own novelty regulation model is to get from non-intelligence to some intelligence by aggregation. Indeed, their is only aggregation of intelligence. It doesn’t seem a satisfying answer.

In another working on bootstrapping I’m transforming the notion of Artificial by Herbert Simon to a more general term. He defines Artificial as human made as opposed to spontaneous, but if your investigating the emerging of intelligence, it should become system made as opposed to spontaneous. A system doesn't need to be intelligent to make other parts, not even for self-replication. In this sense what does artificial actually add as value?

We can understand the artificial by looking how we create laboratories to investigate a specific hidden feature: by creating an artificial condition so that the feature comes observable. In our attempt to understand the hidden feature, we mechanize the artificiality to make the feature more robust and autonomous (for more research on this see Bruno Latour). Indeed, artificial intelligence is exactly what we got with expert systems. They are making the artificial conditions to mechanize some intelligent processes. Ironically it seems very improbable that such research will lead to understanding intelligence itself. As intelligence is aggregated.

It may seem as artificial is quit trivial feature, but we don’t see it that way. Artificial is what makes novelties emerge, not just intelligence. Could we say that the artificial is not important? How would the world look if we only know how things work and not have created technology with it? It wouldn’t be satisfying either. We need both, hence I suggest to create a blur by making AI stand for Artificial-or-Aggregative Intelligence. We may say that our brain is the artifact that creates the condition to aggregate intelligence. I like to make a metaphorical gag with it:

Our minds are like whirlpools, while our brains are like toilets. A day is nothing more than a very long flush. To wake up, we flush our brain, and it stops running until we fall asleep.


Dear Mixel,

interesting perspective!

Your aggregative intelligence reminds me of the 'collective intelligence' concept. Very similar, if not identical. The distinction will depend on the specific definition. Up to you to define that exactly. The distinction with - and relation to - collective intelligence, connected intelligence and cloud intelligence is probably important.

My favorite aspect of intelligence is learning. The major 'human' learning theories have seen an evolution from behaviourism over cognitivism and (social) constructivism towards connectivism. These theories are quite relevant for AI too.
"Intelligence is aggregated" sounds strange for me. Better seems to me "knowledge is aggregated", like we say about human learning that knowledge is something that is constructed (on pre-existing knowledge), in a social way"
Connectivism takes into account that knowledge is something that is connected, and that both knowledge and learning can reside in non-human appliances and organizations.
Intelligence is the ability to make the (distributed) knowledge useful, actionable. Intelligence is only to a very limited amount about aggregating. It's much more about sharing, filtering, synthesizing, sense making, validating, reusing, discussing, decision making, ... and learning!

If you want to go from non-intelligence to intelligence, your aggregating idea is interesting, but the higher aim is learning...

Just my 2 cents, to see if such perspective can add something useful to your nice ideas...

Best regards