Abhinav Saxena

Large language models produce assertion-shaped outputs that often bring their users to true belief. I argue for two conclusions of different strengths: that on the best current evidence, these systems do not transmit testimonial knowledge and that hearers are not now positioned to accord their outputs testimonial uptake, which does not wait on the first. From Burge, Williamson, Goldberg, Faulkner, and Lackey, I derive a Tracking Minimum: a source transmits testimonial knowledge that p only if the truth of p, or the evidential support for p, figures in the explanation of its producing that assertion, and does so in virtue of the source’s own production process rather than of a proxy on which that process depends. It is this endogeneity requirement, not any modal profile, that the five accounts share and current systems fail. The argument engages Burge’s extension of his framework to computational sources, the interpretability literature, and the published case for AI testimony, whose dispute with mine turns on an empirically decidable question, for which an experiment is specified. The positive proposal places LLM outputs within an extended epistemology of instruments, with a result that has gone unremarked: a working barometer satisfies the condition these systems fail.

Within the data-protection-derived legal and governance frameworks that regulate algorithmic systems (the GDPR, the EU AI Act, and the principlist codes that track them), consent and individual authorisation are the load-bearing mechanism of legitimation and redress, and the same frame organises the remedial debate over large-scale algorithmic welfare systems, the Indian Aadhaar architecture being the paradigm case. This paper argues that for a distinctive class of such systems consent is not merely practically strained but conceptually misplaced, and develops a positive alternative. Its central contribution is the concept of constitutive dependence: a condition in which an algorithmic system does not constrain or coerce agents recognised as rights-bearers but produces their operative standing as subjects of recognition and entitlement. Where it obtains, the standpoint of refusal on which consent depends is foreclosed by the system’s algorithmic architecture, which enforces its criteria of recognition without the discretionary judgement by which a human-mediated bureaucracy can recognise the manifestly entitled. Locating the condition within the recognition-theoretic tradition (Honneth, Fraser, Butler; Waelen) as a distinctively infrastructural form of misrecognition, the paper articulates the Constitutive Obligation Principle: where an institution produces its subjects’ operative standing as entitlement-bearers, its legitimacy depends not on their authorisation but on discharging non-waivable obligations of legibility, non-abandonment, and remediation. Developed through the Aadhaar case and offered as a framework for the algorithmic welfare state more broadly, the argument reads Aadhaar not as an outlier but as an early instance of an architectural form toward which algorithmic governance is converging.