Uploaded June 2026 | Updated November 2026, 1 day ago
What happens to human judgment when agentic AI does the thinking first? MIT IDE researchers share new experimental findings on how "beneficial friction" — intentionally designed decision pauses — can protect critical thinking without slowing workers down.
For more information about this research, visit https://ide.mit.edu
In this conference session, Renée Richardson Gosline from MIT IDE, along with MIT IDE postdoc Jean Her, presents research on human-AI decision-making conducted in partnership with Commonwealth Bank of Australia. The session examines when and how people over-rely on AI agents, and what organizations can do about it.
WHAT YOU'LL LEARN
— A simple "playback" intervention — asking workers to articulate their own reasoning before finalizing an AI-assisted decision — reduces over-reliance on agentic AI without adding meaningful time to the task.
— Counterintuitively, having humans decide first and AI second leads to more reliance on the AI, not less — because seeing an AI recommendation puts people in "editor mode."
— When pairing humans with AI models, accuracy alone isn't the right selection criterion; cognitive fit between the human's mental model and the AI's decision logic affects adoption and outcomes.
CHAPTERS
0:00 Introduction
0:53 Renée Richardson Gosline - Beneficial friction: the research framework
23:30 Commonwealth Bank of Australia - Engin on human-centered AI
36:20 Zezhen (Dawn) He: AI alignment with human mental models
FEATURED RESEARCHERS
Renée Richardson Gosline, Research Scientist, MIT Sloan / MIT IDE
Research area: Human-AI decision-making, behavioral science, beneficial friction
Learn more: https://ide.mit.edu/people/renee-richardson-gosline
Zezhen (Dawn) He, Postdoctoral Researcher, MIT IDE
Research area: Personalization in human-AI interaction, mental models
RELATED RESEARCH
Nudge Users to Catch Generative AI Errors
proquest.com/docview/3070032233?fromopenview=true&pq-origsite=gscholar&sourcetype=Scholarly%20Journals
Gathering Intelligence on Artificial Intelligence
https://ide.mit.edu/insights/qa-new-researchers-gather-intelligence-on-artificial-intelligence/
ABOUT MIT IDE
The MIT Initiative on the Digital Economy (IDE) researches how AI, data, and digital technology reshape business, work, and society. Based at MIT Sloan School of Management.
Website: https://ide.mit.edu
#MITIDE #ArtificialIntelligence #HumanCenteredAI
What happens to human judgment when agentic AI does the thinking first? MIT IDE researchers share new experimental findings on how "beneficial friction" — intentionally designed decision pauses — can protect critical thinking without slowing workers down.
For more information about this research, visit https://ide.mit.edu
In this conference session, Renée Richardson Gosline from MIT IDE, along with MIT IDE postdoc Jean Her, presents research on human-AI decision-making conducted in partnership with Commonwealth Bank of Australia. The session examines when and how people over-rely on AI agents, and what organizations can do about it.
WHAT YOU'LL LEARN
— A simple "playback" intervention — asking workers to articulate their own reasoning before finalizing an AI-assisted decision — reduces over-reliance on agentic AI without adding meaningful time to the task.
— Counterintuitively, having humans decide first and AI second leads to more reliance on the AI, not less — because seeing an AI recommendation puts people in "editor mode."
— When pairing humans with AI models, accuracy alone isn't the right selection criterion; cognitive fit between the human's mental model and the AI's decision logic affects adoption and outcomes.
CHAPTERS
0:00 Introduction
0:53 Renée Richardson Gosline - Beneficial friction: the research framework
23:30 Commonwealth Bank of Australia - Engin on human-centered AI
36:20 Zezhen (Dawn) He: AI alignment with human mental models
FEATURED RESEARCHERS
Renée Richardson Gosline, Research Scientist, MIT Sloan / MIT IDE
Research area: Human-AI decision-making, behavioral science, beneficial friction
Learn more: https://ide.mit.edu/people/renee-richardson-gosline
Zezhen (Dawn) He, Postdoctoral Researcher, MIT IDE
Research area: Personalization in human-AI interaction, mental models
RELATED RESEARCH
Nudge Users to Catch Generative AI Errors
proquest.com/docview/3070032233?fromopenview=true&pq-origsite=gscholar&sourcetype=Scholarly%20Journals
Gathering Intelligence on Artificial Intelligence
https://ide.mit.edu/insights/qa-new-researchers-gather-intelligence-on-artificial-intelligence/
ABOUT MIT IDE
The MIT Initiative on the Digital Economy (IDE) researches how AI, data, and digital technology reshape business, work, and society. Based at MIT Sloan School of Management.
Website: https://ide.mit.edu
#MITIDE #ArtificialIntelligence #HumanCenteredAI










