Too bad to be fraud, Midjourney has yet to embark in science.
Cammer, M.
Show abstract
Artificial Intelligence (AI) promises great benefits for data management, for image analysis, and for new discovery, but also creates unprecedented potential for fraud. This brief study explores whether images created by AI pose a threat to the integrity of figures in scientific publications. An attempt was made to use a leading AI image-generation engine, Midjourney, to create plausible facsimiles of images that are commonly depicted in manuscripts in the biological sciences. We conclude that a widely available AI for generated images is not a threat to integrity, yet, but may be in the near future.
Matching journals
The top 1 journal accounts for 50% of the predicted probability mass.
Similar papers in this journal
- Caveolae and scaffold detection from single molecule localization microscopy data using deep learning 93%
- ConfluentFUCCI for fully-automated analysis of cell-cycle progression in a highly dense collective of migrating cells 93%
- Artificial intelligence tool for the study of COVID-19 microdroplet spread across the human diameter and airborne space 92%
Similar papers in this journal
- Tools and methods for high-throughput single-cell imaging with the mother machine 93%
- Identifying images in the biology literature that are problematic for people with a color-vision deficiency 92%
- Automated workflow for the cell cycle analysis of non-adherent and adherent cells using a machine learning approach 91%
Similar papers in this journal
Similar papers in this journal
- Cloud-controlled microscopy enables remote project-based biology education in Latinx communities in the United States and Latin America 94%
- A Comprehensive Framework for Spatio-Temporal Analysis of DNA Damage Foci in Tumor Spheroids. 91%
- Error Rates in SARS-CoV-2 Testing Examined with Bayesian Inference 90%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.