Salesforce AI Ethics Chief: AI Isn't Happening To You, It's Happening For You!
50 min
Salesforce's AI ethics chief Paula Goldman on managing AI at work: why the heaviest AI users feel more in control, not less. Paula Goldman is the Chief Ethical and Humane Use Officer at Salesforce and the author of Manage the Machine: How to Harness Human-AI Collaboration at Work. In this Truth Works episode, host Jessica Neal asks her what AI really does to work. Her answer: small design choices decide whether AI makes people more capable or less, such as engineering scrutiny into an AI agent, letting AI ask questions instead of only giving answers, and deciding in advance where a human must stay in charge. Salesforce's research with AI super users found they felt more agency, not less. WHAT YOU'LL LEARN Does AI make us stop thinking? What cognitive offloading is, and why AI super users feel more agencyWhat habit do the best AI users share? "Engineer scrutiny into the system"Will AI replace people? Why people who delegate well to AI start asking for more human helpHow do you set up an AI agent you can trust? Guardrails designed like the cockpit of a planeWhat are hero use cases, and how do you redesign work instead of throwing AI at people?What should boards measure for AI ROI beyond hours saved?How does Salesforce's HR team map skills to new roles?What do soldiers who held a funeral for their robot teach us about human-AI teams?How do you judge performance when AI agents make mistakes?Where should you never use AI at work?CHAPTERS 0:00 Preview 0:41 Truth Works intro 1:29 Welcome to Truth Works with Jessica Neal 2:51 Salesforce's only Chief Ethical and Humane Use Officer 3:51 From anthropologist and impact investing to AI ethics 5:51 AI at work: the answers are simpler than the headlines 9:29 Will AI make us stop thinking? Cognitive offloading 10:16 AI super users feel more agency, not less 12:31 Delegating to AI agents is the new core skill 16:14 Companies never taught managers to manage 17:59 Hero use cases: redesign the work, not just add AI 19:01 AI agent guard...