← All sources

Philosophical Investigations into AI Alignment: A Wittgensteinian Framework

José Antonio Pérez-Escobar; Deniz Sarikaya · 2024 · Philosophy & Technology 37:80   background medium priority coded

Main argument

Thesis: later Wittgenstein on rule-following, meaning-as-use, and forms of life illuminates the alignment problem - the categories that enable human-human alignment (shared training, practice, use) identify what must be controlled when creating datasets and hard-coded guardrails for AI; casts this in a model of human-human vs human-machine alignment with basic alignment strategies. Implication: rules and principles do not carry their own application (the rule-following considerations), so alignment via explicit constraints inherits Wittgensteinian underdetermination - stability comes from shared practice, not formalization.

Why it matters here

The rule-following version of the alignment problem: Wittgenstein's categories for how humans achieve alignment WITH EACH OTHER (forms of life, meaning as use, rule-following practice) transferred to human-machine alignment - implying rules/guardrails underdetermine application and alignment rests on shared practice. Companion response-piece (2025, to Bangu) also in library.

Reading notes

Compact treatment. Abstract + framing read. The 2025 response to Bangu (PerezEscobar-Sarikaya-2025) treated as an annex - registered no separately coded source; cite pair together.

Pérez-Escobar, J. A., & Sarikaya, D. (2024). Philosophical Investigations into AI Alignment: A Wittgensteinian Framework. Philosophy & Technology, 37, 80.

Close reading — 1 coded units

#1 · pp. 1 · argument
“the later Wittgenstein's philosophy of language and mathematics, substantially focused on rule-following, is relevant to understand and improve on the Artificial Intelligence (AI) alignment problem: his discussions on the categories that influence alignment between humans can inform about the categories that should be controlled to improve on the alignment problem when creating large data sets [...] as well as when introducing hard coded guardrails for AI models.”
VC-THICK

Synthesis-matrix row

supports T3-PROCEDURALISM-INCOMPLETE
rule-following underdetermination beneath the encoding gap

Memos (1)

theoretical · unit #1
The Wittgensteinian rule-following point is the philosophical deep-structure beneath THREE coded findings: S&K's encoding gap (principles don't determine learned policy), Millière's shallow dispositions (trained regularities aren't rule-grasp), and STELA's rule-interpretation problem (annotators must interpret rules via their own background). All are instances of: no formulation carries its own application; application is stabilized by shared practice. For the dissertation this cuts both ways - it deepens the critique of constitution-writing approaches, AND it supports the corpus methodology (folk discourse IS the record of the shared practice within which AI-ethics terms have their use - meaning-as-use makes naturalistic discourse data philosophically load-bearing, not merely sociological). Cite the pair (2024 + 2025 response to Bangu) together.