Cybercrimeology

No News is Bad News: AI Agents, Information Value, Accountability & Democracy

Episode Summary

AI agents are changing how people find, consume and evaluate news. Dr. Aengus Bridgman from McGill University joins us to discuss how AI systems interact with journalism, what happens when news is summarized without clear attribution or compensation and why this matters for the future of the information environment.

Episode Notes

Notes:

• Dr. Bridgman discusses his path from Winnipeg to McGill and how he became involved with the Media Ecosystem Observatory while still a PhD student.
• The conversation turns to the origins of the Media Ecosystem Observatory during the 2019 Canadian federal election and how its work continued through the pandemic, the Hogue Commission and the growing focus on information ecosystem health in Canada.
• Dr. Bridgman explains why COVID-19 and AI have been two major disruptions in the information environment and why AI agents may become one of the main ways people encounter public information.
• The episode looks at the shift from traditional search, where users clicked through to sources, to AI summaries that may give users enough information without sending them to the people or organizations that produced it.
• Dr. Bridgman discusses the problem of value transfer, explaining how aggregators have captured value from original information production and how AI agents may become even more powerful aggregators.
• The conversation considers how common AI news use already is, including the difficulty of measuring it because many people may not recognize that ordinary search now includes AI-generated answers.
• Dr. Bridgman explains what he means by an AI agent: a general intelligence connected to tools that allow it to search, read, summarize and act in digital environments.
• The discussion uses the idea of AI as a “brilliant intern” to explain why these systems can be useful, capable and eager to please, while still lacking judgment about the broader consequences of how they complete a task.
• The episode closes by looking at the harms that may follow if original information production is not sustained, including poorer information, weaker attribution and new challenges for democratic accountability.

About our guest:

Dr. Aengus Bridgman

https://meo.ca/people/aengus-bridgman

https://abridgman.ca/

Papers or resources mentioned in this episode:

Owen, T., & Bridgman, A. (2026). AI News Audit: How AI Models Use and Distribute Canadian Journalism. Media Ecosystem Observatory.

https://meo.ca/work/how-ai-models-use-and-distribute-canadian-journalism

Owen, T., & Bridgman, A. (2026). AI News Audit: AI, Canadian Journalism, and Paths for Policy Action. Centre for Media, Technology and Democracy.

https://www.mediatechdemocracy.com/all-work/ai-canadian-journalism-and-paths-for-policy-action

Other:

Media Ecosystem Observatory

https://meo.ca/

Centre for Media, Technology and Democracy

https://www.mediatechdemocracy.com/