The rise of generative artificial intelligence has destabilized journalism in familiar ways. The transportation model looks fragile. Copyright is disputed. The marginal cost of producing words has collapsed. It's tempting to conclude that news is becoming less valuable, just as machines are becoming more fluent.
So far, the opposite has happened in Mongabay.
AI systems are good at rearranging existing information. They quickly summarize, paraphrase, and answer questions. What they cannot do is observe the world directly, make responsible judgments, and create new facts. These limitations are no coincidence. They are embedded and put journalism in a more central position than before.
This became clear in an unexpected way. At Mongabay, a nonprofit newsroom focused on environmental reporting, we have chosen not to block access to our work by generative AI systems. Many publishers do the opposite, citing copyright concerns, energy usage, or disintermediation concerns. Our reasoning was realistic. We already allow other outlets to republish our reports as part of our impact strategy. If AI tools are going to answer questions about forests, fisheries, and biodiversity, the answers seem better served by reported journalism than by more detailed sources.
I thought this would reduce traffic. If people can get what they need from an AI interface, why would they click through?
That didn't happen. ChatGPT has become one of the largest traffic sources outside of Google. What's even more surprising is that readers who access us via ChatGPT spend significantly more time with our articles than readers who access us from other platforms. They check the sources, peruse them, and seem to stay.
This isn't because AI suddenly cares about journalism. Maybe it's because the users are doing it. When answers are rich and frictionless, provenance begins to matter. Readers seem to want to know where the claims come from and whether someone is responsible for them.
That accountability is a core asset of journalism in the age of AI. Newsrooms verify events, document evidence, and attach names and institutions to what they publish. In a world of synthetic text, images, and video, validation is no longer just an internal process. That's what readers are actually looking for.
Judgment will also become more visible. AI systems leverage patterns in historical data to optimize likelihood and relevance. It is journalism that decides what is worth attention in the first place. We choose which harms to investigate, which voices to elevate, and which silences are unacceptable. These choices reflect values and public responsibility. Currently, they cannot be inferred statistically.
For the same reason, initial reporting is also important. Investigations, field reports, and beat expertise generate information that did not previously exist. AI systems rely on such tasks, even as they obscure their dependencies. Without a steady supply of new reports, the model cannot learn anything new beyond synthetic data.
Meanwhile, trust is becoming increasingly scarce. Reliable news organizations explain how they know what they know and correct their mistakes in public. AI systems are largely opaque. As uncertainty increases, transparency becomes a competitive advantage.
This does not mean that journalism is safe from disruption. AI will reshape workflow, distribution, and revenue. But for now, it seems to be helping clarify the role of news, rather than erasing it.
When there is an overabundance of language, it becomes difficult to find the truth. As a result, the need for journalism and the institutions that practice it is growing, not diminishing.
