By Peter Wilmot, Chief Product Officer at Shopsense
For years, publishers have watched their content power the internet with diminishing returns. Search engines summarized their reporting, social platforms siphoned audiences, and ad-tech middlemen captured most of the revenue.
Publishers created the raw material, and everyone else captured the value.
Generative AI has escalated that long-running business imbalance into a full-blown crisis. More than just referencing the news, AI absorbs it, rewrites it, and delivers answers directly to users. When AI-generated summaries obviate original reporting, the economics of journalism collapse.
Publishers have split into two camps: those who license, and those who sue.
The New York Times’ lawsuit against Perplexity alleges widespread scraping and reuse of its journalism, attempting to create a legal moat around original reporting. Meanwhile, platforms like Meta are striking paid licensing deals with publishers to train and enhance their AI systems.
Authoritative content is a valuable, licensable asset. AI developments have catalyzed action.
Publishers ought not view these deals as their salvation. AI licensing helps, but it doesn’t solve everything. And it brings serious risks.
For now, platforms still need publishers
AI companies are learning that building trustworthy systems not only requires massive data scraping, but verified facts, editorial judgment, and real-time reporting. Hallucinations, misinformation, and unsafe outputs become reputational and regulatory liabilities.
AI companies can’t just scrape publisher content once and be done. Publishers possess key qualities that give them leverage, including:
- Fact-checked journalism
- Archival depth
- Specialized analysis
- Reporting from places algorithms can’t reach
But publishers must not imagine this dependence will last forever. Because a one-time scrape becomes stale so quickly, AI companies are racing toward a world where they won’t need publishers. Synthetic data, model self-training, user feedback loops, and proprietary sensors could eventually reduce reliance on outside content. Once AI companies fully control the “looking glass” through which people discover information, they’ll feel a good deal less compelled to pay the creators of that information.
Platforms Have a Long Record of Bait-and-Switch
Publishers have lived this cycle many times over with Google AMP, Facebook Instant Articles, and the 2018 Facebook “pivot to video.” Publishers have been burned repeatedly by initiatives that promised traffic and revenue but were quietly abandoned not long after the publishers restructured their businesses around those initiatives.
Platform incentives and publisher incentives rarely align for long.
So why should AI licensing be any different? What happens when AI platforms no longer need external data?
Publishers Cannot Replace Journalists With AI
Some publishers are tempted to reduce newsroom costs by generating content with AI. That, too, is a trap.
When publishers flood the ecosystem with low-quality, machine-written content, their licensing value plummets, their differentiation disappears, their brand trust erodes, and AI companies look elsewhere for authoritative sources.
AI-generated filler may boost short-term output, but it destroys long-term leverage. What AI companies want, and will pay for, is original reporting they cannot produce themselves. Newsrooms must invest in human journalism to increase their value in the AI economy.
Publishers Need New Revenue Channels
Licensing alone won’t restore publisher economics. At best, it’s supplemental income. To rebuild sustainable businesses, publishers need to diversify in ways that deepen audience relationships, create proprietary value, and insulate them from platform volatility
They will also need to build:
- Direct audience revenue: subscriptions, memberships, and paid communities that create predictable, defensible income
- Commerce-driven content: service journalism that connects discovery to purchase in ways AI intermediaries can’t fully replicate
- Branded editorial franchises: tentpole series, events, and IP that can live across formats, platforms, and revenue lines
- Retail and affiliate partnerships: integrated shopping experiences built around trust and expertise
- AI-powered tools: workflow accelerators, research assistants, and summarization tools that enhance newsroom productivity without replacing reporting
The main opportunity lies in rethinking how content itself becomes a product.
The Stakes Are Bigger Than the Industry
If the economics of journalism fail, society loses the fact-checking layer on which the entire information ecosystem, and therefore our democracy, rely.
AI models trained on unverified data will drift further from reality, misinformation will compound, and citizens will find themselves in a world where knowledge is generated, validated, and distributed by the companies that control the discovery interface.
Healthy AI requires healthy publishers. Otherwise the entire chain of knowledge breaks.
