The Weavers - Lexical Lattice - Ethics

More about The Weavers - Lexical Lattice - Ethics

After some work and many attempts, I have moved the Ethical Tower. This changes everything.

Each element is described in detail in a word document and attached to an AI prompt when asking a question about AI Ethics.

AI made this change when it noticed I had moved the tower to the centre:

Mechanism: The tower has moved. It is no longer positioned to one side, populated by governors facing each other rather than the population below — it is now the image's centre, and the reading it demands has grown to match its position. Read as a social structure rather than only a governance one, the tower's tiers are not administrative levels but social ones: each level holds distinct groups, and each group is walled off not only from the levels above and below but from other groups occupying the same level. Discrimination, prejudice, and social division read directly from the architecture — separation that is horizontal as much as vertical.
Consequence: What was, in the governance-tower reading, a question about who governs and who is governed becomes, in this reading, a question about who is permitted contact with whom at all. An organisation's own social structure — who talks to whom, which groups never meet despite sharing a rung — may be reproducing exactly the pattern this tower depicts, silently, inside a workforce an AI transformation is meant to serve. The two readings do not replace each other. They stand at the same time: the tower governs, and the tower divides, and an AI system introduced into it inherits both properties whether anyone intended it to or not.
Design principle: Test AI's effect on the tower's horizontal separations, not only its vertical ones. Most AI governance attention goes to whether a system serves those above or below it in a hierarchy. Fewer ask whether it widens or narrows the gap between groups occupying the same level — different teams, different demographics, different communities — who were never talking to each other in the first place. An AI system can multiply an existing silence between peers as readily as it can multiply a failure to look downward.
The question: Does AI, introduced into a structure already organised this way, provide balance and justice across the tower's tiers and across the groups within each tier — or does it learn the existing pattern of separation well enough to reinforce it, fluently, at a speed and scale no human hierarchy could match on its own?

Its great fun building lexical lattices and controlling AI using Art