I distinctly remember the day I saw a western blot band stretched, rotated, and pasted into another panel. Zoomed out, it looked like a perfectly normal blot; the imposter band sat amongst the others like it had always been there.
Sitting at a long table with the other graduate students on my training grant, I watched as our professor showed us example after example of images from published scientific papers that had been manipulated to embellish the data. I really appreciate that course and the other research integrity courses I took during my research training for teaching me and my peers how to spot bad science and what to do when we encounter it. It made me a better scientist when I was in the lab, and now, it makes me a better journalist.
When bad science infiltrates the publication record, researchers unwittingly build their own research programs around shaky science. Not only does this waste researchers’ time and money, but it affects real people’s lives. When allegations of misconduct began swirling around the basic science that led to the development of Cassava Science’s Alzheimer’s disease drug simufilam, the treatment was already in clinical trials. Now, five of those papers have been retracted, and while the drug is in Phase 3 trials, many scientists are waiting to see how the results pan out with the drug’s mechanism of action now under question.
In recent years, technologies such as ChatGPT have emerged that take scientific misconduct to a whole new level. In fact, the popular blog Retraction Watch now has an entire section on its website devoted to papers and peer review reports that have evidence of ChatGPT use, and the documents featured there are likely only a fraction of the ones out there. But even so, somewhat “old fashioned” examples of misconduct like plagiarism, improper copying and pasting, and plain old lying are still rampant in the scientific literature.
While artificial intelligence tools may have upped the ante for improper conduct, they’ve also made it easier for those who are looking for misconduct to spot it. More instances of research misconduct are being reported and making news headlines every year. Prolific research misconduct sleuth, Elizabeth Bik has practically become a household name for her ability to spot shoddy science, and newcomers such as Sholto David uncover more suspect papers as well. But they can’t catch everything.
Preventing research misconduct starts before someone collects even a single piece of data. It begins with laboratory leaders who foster an honest research environment and encourage openness among all those who work there. Research integrity courses like the ones I took in graduate school help, but finding ways to change the “publish or perish” culture rampant in academic research will likely benefit science the most. A couple ways to do that would be to incentivize replication studies and research ethics training. Systemic changes that make doing ethical research easier will improve outcomes for everyone from the researchers doing fundamental research that will lead to new drugs to those who will benefit from the treatments.
