Algorithmic investment literacy: Teaching students to evaluate, calibrate, and ethically use AI investment advice

Abstract

The diffusion of robo-advisors and generative artificial intelligence has changed the central problem of investment education. When advanced language models score near the ceiling on standard financial-literacy tests, the decisive competency is no longer whether learners possess financial knowledge, but whether they can judge the machine-generated advice they receive. This paper defines Algorithmic Investment Literacy (AIL) by a single operation that no neighbouring construct centres: the learner lets a personal-suitability judgement, whether a recommendation fits this investor's own goals, risk tolerance, horizon and constraints, govern reliance on an adviser whose reasoning is opaque, so that suitability can veto a reliable-looking recommendation and license a plainly generated one. The human-automation reliance literature calibrates reliance to a system's verifiable task accuracy rather than to a recommendation's private, unverifiable fit with the person; financial literacy concerns what a person knows and not how they appraise an external adviser; generic AI literacy asks whether an output is sound in itself and not whether it is right for this person. Each supplies part of the operation; AIL joins them, and we argue it predicts reliance-decision quality beyond their additive combination, showing construct by construct the behaviour AIL predicts that each neighbour cannot. Four finance-specific dimensions follow from the operation rather than standing beside it: provenance awareness, suitability appraisal, suitability-governed reliance, and fiduciary and responsible use. The paper advances nine propositions linking pedagogical and learner antecedents to AIL and its outcomes, and translates the construct into a learning progression, a worked classroom task with an assessment rubric, and a research agenda, giving educators a teachable, assessable target and researchers a foundation for instrument development and validation.