Engineering Trustworthy AI: A Developer Guide for Empirical Risk Minimization

TL;DR

This paper provides a developer guide for designing trustworthy AI systems using empirical risk minimization, addressing issues like bias and opacity to meet emerging trust standards.

Contribution

It translates key trustworthiness requirements into practical design choices for ERM-based AI systems, offering actionable guidance.

Findings

Guidelines for incorporating trustworthiness into ERM

Strategies to reduce bias and improve transparency

Framework aligning ERM with trust standards

Abstract

AI systems increasingly shape critical decisions across personal and societal domains. While empirical risk minimization (ERM) drives much of the AI success, it typically prioritizes accuracy over trustworthiness, often resulting in biases, opacity, and other adverse effects. This paper discusses how key requirements for trustworthy AI can be translated into design choices for the components of ERM. We hope to provide actionable guidance for building AI systems that meet emerging standards for trustworthiness of AI.