BOSTON - BostonChron -- Somewhere today an AI tool ranked a stack of job applicants and sent a hiring manager the top five, already sorted. The ranking criteria came from somewhere else, and nobody who will use that list can explain the criteria. The decision happened, and it never felt like one.

Kevin Hanegan calls this the quiet handoff. Judgment moves from people to systems one reasonable convenience at a time, and nobody signs anything. His new book, Who Decides? A Field Guide to Human Judgment in Artificial Intelligence, is not about the room where models get built. It is about the room next door, where a manager decides whether to act on what the model just said.

"The machine is sometimes the fairer, sharper, more reliable judge, and refusing to use it is its own kind of failure," Hanegan says. "It is also, often, confidently wrong in expensive ways. The skill that matters now is not building AI. It is knowing which decisions belong to the machine, which belong to the human, and where the two need to check each other."

More on Boston Chron

• Commercial Sales Launch Following FDA Approval; National Sales Force Now in Action as January Shipments and Major Market Opportunity Take Shape
• ViewDNS.info Launches Monitors, Alerting Teams the Moment Something Changes on Their Domains
• Philadelphia-Area Entrepreneur Nouho M Kanneh Builds $37 Million Net Worth Across Real Estate, E-Commerce and Retail
• Gray Code Entertainment Closes the Still 7 Series as HoodBoss Drops 'The Last 7' for Just $1
• Get Ahead of It: Testology Inc. Highlights a New Approach to Multi-Cancer Early Detection

Who Decides? A Field Guide to Human Judgment in Artificial Intelligence (ISBN 978-1629673479) grounds that argument in cases where uncalibrated trust produced public failures Underneath these failures, Hanegan finds one recurring pattern: the proxy trap. You can rarely measure what you actually care about, so you count something that stands in for it. A widely used healthcare algorithm ranked patients for extra care using what had been spent on them, on the reasonable theory that sicker people cost more.

Hanegan also introduces the Automation Trust Check, five questions covering proxy, incentive, recourse, reliability, and stakes, designed to run in a single meeting before a system goes live. The book also challenges the standard governance answer. Keeping a human in the loop, Hanegan argues, tends to decay into a person initialing whatever a fluent machine drafts. The alternative he proposes puts the machine in the loop, with the person deciding and the model drafting.

More on Boston Chron

• City of Boston Celebrates Grand Opening of New EMS Station on South Boston's Waterfront
• Contracting Resources Group Named a 2027 Elev8 GovCon Honoree by OrangeSlices
• SLYNUMBER Gives Consumers More Confidence in Financial Texts with Verified Origination
• The World's No.1 Superstar® Brings Christmas Back in Grand Style with Music, Cinematic Glamour, and Luxury Fashion
• New Six-Figure Media Agreement Builds Powerful Momentum, Expanding Government Contracts, FBI Access and Growing AI Recognition: Cyabra (NAS DAQ: CYAB)

ISBN: 978-1629673479
Paperback: $9.99 / Kindle $4.95
Worldwide Distribution: Ingram

About the Author
Kevin Hanegan is a data literacy expert who helps organizations decide how much to trust their machines. Rather than building AI systems, he works with the leaders responsible for deciding when those systems should be trusted, challenged, or overruled. He is the founder of DataIntoWisdom.com and the author of Turning Data Into Wisdom, Data Literacy in Practice, and the Data Literacy Fundamentals series available on Amazon.