Policy

AI-use policy

AI systems may assist with bounded research organization, comparison, testing, or draft development. They do not independently approve or publish Token & Taste material.

Human responsibility

A human editor remains responsible for source selection, factual verification, interpretation, privacy decisions, editing, and publication. Generated citations and product claims are checked against the underlying source or direct observation.

Evidence and disclosure

Generated examples are not represented as interviews, measurements, firsthand observations, or real experiment results. Meaningful AI assistance is disclosed in article metadata. Confidential or personal material is not supplied to an external AI service without authority and an appropriate data-handling review.

Limits

AI output can be incomplete, biased, outdated, or fabricated. Fluency is not treated as evidence. Readers should be able to distinguish supplied evidence, model output, and the editor's conclusion.

Effective 11 August 2026. Owner and professional legal review remain required before public launch.