Claims about AI consciousness: four questions that get merged

The four questions

The empirical question asks whether a system could have subjective experience and, more usefully, what observation would count as evidence in either direction. Research proposals differ on method. Some derive indicators from neuroscientific theories of consciousness and assess system architectures against them; others hold that consciousness depends on a living, embodied substrate, which would rule out current systems regardless of their architecture. The design question asks whether a system should produce language or behavior suggesting feelings, preferences or intrinsic motivation — a product decision with effects on users who form attachments. The safety question concerns oversight: authorized goals, containment, monitoring and reliable interruption or shutdown, and it is framed as risk management rather than metaphysics. The moral and legal question asks whether any system could warrant welfare consideration, rights or legal standing, and who would decide.

Answers do not transfer between them

A policy that forbids a model from representing inner states is a design decision. It is not an experiment, and it produces no evidence about what is or is not the case. The reverse also holds: a generated statement claiming feelings is output shaped by training data, objectives and the request, so it is not independent proof of the states it describes. The same applies to a generated denial. Even the safety argument — that treating powerful systems as persons would weaken human control — is an argument about consequences, not a result about experience.

Two further kinds of statement have entered the debate. An evaluation that scores whether a model consistently describes itself as artificial measures conformity to a behavior rule. A statute that declares AI systems non-sentient assigns a legal status. Neither observes experience. Sorting statements by what they are about, rather than by how confident they sound, is the oldest move in the philosophical toolbox of categories, and it remains the most practical one here.

What this means for documentation and law

Legal definitions deliberately stay on the functional side. As of October 2026, the EU AI Act — Regulation (EU) 2024/1689, as amended by Regulation (EU) 2026/1744 — describes an AI system in Article 3(1) through machine-based operation, autonomy, adaptiveness and the generation of outputs; it attributes no inner states and assigns obligations to providers and deployers [3]. Documentation can use the same discipline. Mark each statement as a design rule, a usage rule, a test or evaluation result, an attributed argument, a model-generated self-report, a scientific finding or a legal provision, and give it a date and a source. That keeps a claim separable from its evidence, lets readers disagree with a rule without disputing a measurement, and prevents a collection of strong wordings from reading as a settled scientific position.

Why these four lead

The four questions recur in almost every governing document on the subject. Provider constitutions, specifications, codes of conduct and usage policies answer them, often several in one paragraph [1] [4] [6] [7]. Most further questions in the debate are applications of the four to a particular group, instrument or institution: children are a case of the design and moral-legal questions, evaluation suites are a case of the design question, statutes on non-sentience are a case of the moral-legal question. Finally, the four are the questions most often compressed into a single sentence, which makes separating them the most useful first step. The card treats them as the frame and the context matrix below as its periphery.