論文偽造の阻止

Scientific fakery is a centuries old problem. Twinned with the long history of hard-working scientists earning fame for genuine discoveries, runs a tawdry history of those who were willing fabricate results to falsely gain prestige. Fraud in the past relied on bespoke fakery, but today's fraudsters can exploit the online scientific world to quickly create realistic looking papers on an industrial scale. Fraudsters are using open data sets to create meaningless analyses and combining these results with text from large language models. There has been an explosion of these low value papers using openly available and highly regarded data sets, such as the US National Health and Nutrition Examination Survey (NHANES). The paper miners will likely exploit whatever open data resources they can find until data custodians put more stringent controls in place, or journals and publishers push back. Some scientific data may be too open, even though making research data openly available is a recommended policy for increasing research integrity. Journals and researchers need to be aware of this new threat to research integrity.