AI Plagiarism and Integrity in Science News
Key Developments in Research Integrity
This week, the world of academic publishing saw several significant events. A cardiology researcher in Italy has accumulated 17 retractions, with most cases involving duplication. In a striking example of publication delays, some authors who requested Elsevier retract their papers in 2012 are still waiting for action on one article. Meanwhile, the collective effort of online sleuths, described as 'Hive mindfulness,' led to the retraction of a paper on social connection. In another high-profile case, a top education researcher is now facing a court battle and a university review over plagiarism claims. Additionally, the journal PLOS One has issued expressions of concern for four papers due to issues with overlapping control data.
The Growing Influence of AI in Research
The intersection of artificial intelligence and academic integrity is a major point of discussion. A significant new study reveals that ChatGPT tends to ignore retractions of scientific papers, potentially spreading discredited information. This highlights the growing concern over what constitutes plagiarism in an era where AI-generated papers pose new risks. On the other hand, AI is also being positioned as a solution, with new research exploring its use for detecting ethical breaches and misconduct in manuscripts. One study even found that AI can be better than human experts at distinguishing between human-written and AI-written papers on stroke research. However, the current state of AI writing disclosures is considered a joke, with calls for significant improvements.
Broader Challenges in Academic Publishing
Beyond AI, the academic community continues to grapple with systemic issues. Researchers have proposed a new approach to peer review, using the ‘Swiss Cheese Model’ to better flag problematic manuscripts. The peer review system itself is described by some as creaking and on the verge of collapse.
Concerns about research quality and fraud are mounting:
- An investigation highlighted how bogus research papers are damaging science, fueling a corrupt industry.
- A survey in nuclear medicine research found that nearly 40% of scientists witnessed colleagues participating in misconduct.
- The problem is so widespread that some say scientific fraud has become an industry that is growing faster than legitimate journals can handle.
There are also ongoing debates about publishing models. Some are calling to end unchecked publishing fees and unpaid peer review, while others suggest a radical one-a-year publication limit for researchers. Many express fears that essential tools like Google Scholar are becoming unreliable, a change for which academia is not prepared.
