Imagine a world where entry-level journalism jobs have been outsourced to AI, and the industry has experienced what AI expert Alex Mahadevan calls a “market failure” disrupting the “talent pipeline.”
Mahadevan visited Harvard Extension School’s journalism ethics class on March 26, 2026 to discuss AI’s influence on journalism and his hypothesis for the landscape of the future. Mahadevan’s work as a “media literacy trainer” and director of MediaWise at the Poynter Institute leads him to believe that journalists need to familiarize themselves with AI tools, but he also notes several cautions for the days ahead.
Mahadevan warned that integrating AI tools with search engines like Google is influencing newsroom profits. When individuals can quickly read a bite-sized AI summary, Mahadevan said, they’re less likely to click through to read the news. Mahadevan described search-generated “traffic” as “totally off a cliff” or potentially almost non-existent in the near future.
But the threats AI poses to the journalistic process are not only financial; Mahadevan fears that the ethics of the profession are under fire when AI is at play. He said AI tools often offer the opposite of core journalistic values like independence and accuracy. Mahadevan warned that journalists must develop “guardrails centered on humans and verification,” estimating that only 50% of newsrooms have formally encoded their journalism ethics.
Mahadevan expressed concern that people will become accustomed to AI’s “sycophantic” rhetoric. He gave an example of asking an AI agent if he should go for a run; he didn’t really want to, and the AI agent picked up on that and offered him reasons his instinct was right.
And there are other challenges in the age of AI. Alex explained the “disclosure paradox” to students, saying that if journalists do not disclose that they are using AI, their audiences doubt them, but the more news organizations do disclose, the more they are doubted as well. “Audience trust is sacred to us,” Mahdevan argued; “it is our currency.”
Mahadevan posited that AI may function not only as a source of disinformation — “we have to verify . . . every single photo that comes out of this administration,” he explained — but also as a scapegoat for the dismissal of damning evidence. For example, he expounds, what if the Access Hollywood tape came along today, and politicians could dismiss it as made up by AI? People are moving, Mahadevan says, from “skeptical to cynical.”
At the same time, Mahadevan sees processes such as “vibe coding” as the “biggest opportunity” for journalism today (though he notes it comes with risks as well). Vibe coding, Mahadevan explained, is when journalists ask an AI agent — using ordinary conversational speech — to produce a tool that not long ago they would have needed coding expertise to develop. Thus, Mahadevan said, “any journalist has the ability to be a data journalist.”
It is “on us as journalists to raise our voices,” Mahadevan told the class. Students closed the night by exchanging positive feedback on his talk and pondering the future amongst themselves.