Bacterial Wisdom As Template for Artificial Free Will
If any genuine “free will” exists, it is at the level of the “I-ness” of a system, the decision making routine, that it comes into play. Before we dive into the technicalities of this issue, let’s first try to brainstorm on what can be understood by “free will”. Although intuitively we “know” what “free will” is, just as we know what consciousness is,it is extremely hard to define it in words. Let’s try to build an ontology “free will” by reciting its features and by drawing the borders of this concept from the notions of what it is not.
I followed a very interesting discussion on the issue of free will and whether it is needed in AI, which I will neither repeat nor summarise here, but a number of striking concepts of which I will use in this essay. I do not claim to have come up with those concepts myself nor do I claim to be an expert on the issue, but I believe that I can add some interesting concepts to the discussion deriving from Ben Jacob’s “Bacterial Wisdom”, “Global Brains” and “Societies-of-Minds”. I will also propose to incorporate an artificial functional mimic of “Free Will” in a Webmind such as the AWWWARENet (Artificial World Wide Web Awareness Resource Engine Net).
A number of concepts stood out above the noise of the aforementioned discussion, which I’ll mention here as features (and non-features) of the “free will ontology”:
“Choice, override, randomness, unpredictability, (non)determination, chaotic, (non)causality and evolution”.
Indeed, for a “Will” or decision-taking routine to be “free”, it must be able to override those possible decisions, which are “causality-determined”. In Goertzel´s Webmind the discriminating faculty is the AttentionBroker routine). In the AWWWARENet, the AttentionBroker presents its conclusions, what course of action is to be taken as being the most rational, as having the highest probability of success, to the I.I.I (Identity,Initiative and Illusion generating routine). In as far as the system has an “override” function, the system appears to be endowed with a faculty of “choice” to an outside observer of the system.
The need for a random-picking faculty arises, when the AttentionBroker present the I.I.I-routine with more than one equally likely options i.e. options with identical priorities.
The issue becomes more poignant, when due to a scarcity of resources or time imposed resource constraints not all options can be carried out simultaneously or worse are mutually exclusive i.e. some must be sacrificed at the expense of others.






