Vilas Dhar: Journalism has yet to begin its AI revolution

Real AI integration requires investment, specialized talent, and workflow upheavals

By Vilas Dhar

February 4, 2026

This essay is part of a series on the 2026 Lenfest Institute-Aspen Digital Local News Summit, an annual convening of the country’s leading journalists, publishers, funders, news creators, and other industry professionals. 

Finance now executes 70% of equity trades through AI algorithms. U.S. hospitals have doubled their adoption of AI in two years. Amazon deploys over 1 million warehouse robots that have boosted productivity per worker by more than 20 times in under a decade. These interventions represent wholesale AI-enabled redesigns of how entire industries operate, from workflows to workforces to strategies.

Journalism, meanwhile, is still testing the waters.

There are pilots, of course. The Associated Press automates corporate earnings briefs. The Washington Post built Heliograf to cover simple election results. Reporters dabble with ChatGPT for research and drafting. But the typical newsroom of 2025 still runs on the same roles, routines, and deadlines as 2015, with a thin veneer of AI tools at the margins.

The numbers confirm this glacial pace of adoption. In a recent survey spanning more than 70 countries, nearly 80% of newsrooms had no formal AI policy. Another study found that while publishers claim AI is strategically important, only 13% of their employees can identify any clear AI strategy from leadership.

Why the paralysis? Legacy media companies, battered by earlier tech disruptions, now often operate under tight economics and cautious cultures that prize survival over risk. Real AI integration requires investment, specialized talent, and workflow upheavals. So for publishers, incrementalism wins the day, and organizations settle for using AI tooling for efficiency rather than for reimagining how news gets made, consumed, and shared.

How does the industry break out of this self-imposed stasis?

The first step is to stop treating AI as a collection of shiny tools and start treating it as a catalyst for redesigning how journalism actually works. That means asking harder questions: What do we want journalism to achieve in an AI-saturated world, and what won’t we sacrifice? If we built a newsroom from scratch today, how would we integrate AI into every step while upholding editorial standards? How do we redefine roles so that human judgment and machine capability combine effectively?

Most of all, the industry needs to ask a different question entirely. Instead of asking “Which new AI tool should we try?”, newsroom leaders should ask “What systemic change would make us more effective storytellers?” That might mean radically different newsroom structures, teams that mix developers and journalists by design, or workflows that automate routine information so humans can focus on high-impact investigative work. It means confronting sacred assumptions: Do we need to produce as many commodity updates, or could machines handle those while journalists double down on original reporting? Can we continue the 24/7 news cycle as it exists, or should we use automation to reclaim time for the context and analysis that readers actually need?

Journalism’s core values, truth, accountability, and serving the public, will not survive by avoiding change. They survive through adaptation, which requires leadership willing to overhaul venerable institutions and build systems that amplify accuracy, trust, and emotional resonance.

The storytellers cannot afford to sleep through the revolution happening in our midst.