The No. 1 Question to Ask When Evaluating AI Tools

Name: The No. 1 Question to Ask When Evaluating AI Tools (English version)
SKU: SMR64314

Sarah Lebovitz; Hila Lifshitz-Assaf; Natalia Levina

Product #:SMR64314

Supplier:MIT Sloan Management Review

Discipline:Management Science

Details:Article (PDF) • 8 pages • 2023-04-01

Subjects:

Analytics and data science Data science

Your Price:

$10.54

Purchase this material with a free Educator Account and enjoy more benefits:

Free Educator Copy
Free Teaching Notes
Build your coursepack
Discounted academic price

Developers of many new AI solutions produce statistics showing that the tools make critical decisions with greater accuracy and efficiency than humans. But managers tasked with evaluating these applications need to peel back the layers of developers’ performance claims and focus on the ground truth used to train and validate the AI tool. This article explains what ground truth is, how to identify it for a given tool, and how to understand the implications for tool quality.