English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The problem of human trust in artificial intelligence is one of the mostfundamental problems in applied machine learning. Our processes for evaluatingAI trustworthiness have substantial ramifications for ML's impact on science,health, and humanity, yet confusion surrounds foundational concepts. What doesit mean to trust an AI, and how do humans assess AI trustworthiness? What arethe mechanisms for building trustworthy AI? And what is the role ofinterpretable ML in trust? Here, we draw from statistical learning theory andsociological lenses on human-automation trust to motivate an AI-as-toolframework, which distinguishes human-AI trust from human-AI-human trust.Evaluating an AI's contractual trustworthiness involves predicting future modelbehavior using behavior certificates (BCs) that aggregate behavioral evidencefrom diverse sources including empirical out-of-distribution and out-of-taskevaluation and theoretical proofs linking model architecture to behavior. Weclarify the role of interpretability in trust with a ladder of model access.Interpretability (level 3) is not necessary or even sufficient for trust, whilethe ability to run a black-box model at-will (level 2) is necessary andsufficient. While interpretability can offer benefits for trust, it can alsoincur costs. We clarify ways interpretability can contribute to trust, whilequestioning the perceived centrality of interpretability to trust in populardiscourse. How can we empower people with tools to evaluate trust? Instead oftrying to understand how a model works, we argue for understanding how a modelbehaves. Instead of opening up black boxes, we should create more behaviorcertificates that are more correct, relevant, and understandable. We discusshow to build trusted and trustworthy AI responsibly.

Trust in AI: Interpretability is not necessary or sufficient, while black-box interaction is necessary and sufficient