Sociological Bias in AI Development

Researcher preferences and goals of mimicking human reasoning shape model development, potentially causing convergence toward human-like representations

Neighborhood — ranked by edge-count

Concepts (1)

concept

Representational Convergence
supports
The central empirical phenomenon: different neural networks trained on different data/objectives develop increasingly similar representations

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Simplicity Biasconcept0.735
The tendency of deep networks to implicitly favor simpler solutions that fit the data, driving convergence
AI alignmentconcept0.724
Field within which this work has implications for evaluating alignment progress.
Many recent discussions of AI, and its impact on individuals and on society, are importantly incomplete because they neglect Diverse Intelligence, synthetic morphology, and developmental biology.claim0.723
Central thesis of the paper — the framing premise from which all other arguments follow
Patterns in AI self-reports should be compared across different models to identify structural commonalities.claim0.721
How Should The Development Of Artificial General Intelligencequestion0.719
Inductive Biasconcept0.719
Assumptions or preferences (e.g., parsimony) that determine how a learning system generalizes beyond training data
There are no obvious technical barriers to building AI systems which satisfy the indicator properties.claim0.718
Feasibility claim about near-term conscious AI.
How does the individuation of AI systems affect consciousness attributions?question0.718
Open question about copying and distributed systems.