Three Dangerous Lines in Research Communication
Most communication risks fall into three groups
Many risks in academic communication can be grouped into three categories: fake data, fake experts, and fake authority. They look different, but they do similar work. Each makes readers believe the author has stronger evidence than is actually available.
Among the three, bought citations deserve special attention because they disguise slow academic feedback as if it has already happened.
Fake data is not always an invented number
Fake data does not always appear as a number invented from nothing. More often, a small sample is written up as a general rule. An internal estimate is described like a measured result. A vendor dataset is used without examining the method and then placed directly into the conclusion.
The danger is that the communication layer removes the uncertainty that should have traveled with the evidence.
Fake experts are becoming easier to manufacture
Fake experts are also becoming easier to create. AI can generate fluent expert commentary and plausible author biographies. But expertise ultimately depends on concrete experience. A person who works on image analysis can speak to image-analysis problems, but that does not make them responsible for every conclusion in a study. A collaborator can discuss one experiment without guaranteeing every sample.
When a quote or profile borrows expert authority too broadly, it misleads readers about what has actually been validated.
Fake authority hides inside familiar formats
Fake authority is more subtle. Paid publication can look like a normal paper. Guest authorship can look like collaboration. Awards and rankings can be described more strongly than their selection process deserves.
Bought citations are even more dangerous. They take academic recognition, something that should accumulate through real use and discussion, and make it look as if that recognition has already arrived.
Why citation manipulation spreads damage
Readers naturally treat a heavily cited paper as more authoritative. If those citations come from transactions, mutual-citation circles, or batch arrangements, they no longer say much about quality.
Citations enter author profiles, journal metrics, project materials, and institutional descriptions. Later readers often see only the elevated number, not the questionable path that produced it. Once citation manipulation is detected, the damage reaches far beyond a single article.
Reference material: U.S. Department of Justice announcement on the Dana-Farber settlement; Retraction Watch on the Op4G fake-data case; SciIntBench preprint; Scientific American on large language models and academic fraud.