Connect with us


Artificial intelligence may speed up scientific fraud alongside advancements



AI could accelerate scientific fraud as well as progress

The aim of a recent session organized by the Royal Society in partnership with Humane Intelligence was to challenge AI models and test their ability to produce fake academic articles. Some results were found to be misleading, such as claims that ducks could indicate air quality or lavender oil could treat long covid. However, the real concern lies in the manipulation of AI for academic misconduct and fraud.

Researchers like Guillaume Cabanac have discovered instances where AI-generated phrases have been accidentally included in academic papers, indicating the use of LLMs without proper disclosure. This deceitful practice has led to a rise in research-integrity cases, with publishers like Taylor and Francis investigating a significant increase in fraudulent activities.

Not only are academic papers being doctored with the help of AI, but images in scientific papers are also being manipulated. Microbiologist Elisabeth Bik has identified numerous papers with identical features in images, suspecting them to be AI-generated to support false claims. The challenge lies in identifying machine-generated content, as current methods are not foolproof.

Furthermore, there are concerns about AI models being trained on outdated data, leading to potential inaccuracies in scientific discovery. The issue of model collapse, where AI churns out illegible or erroneous outputs, poses a significant obstacle in using AI for research. The lack of transparency in AI algorithms and the difficulty in understanding their inner workings raise questions about the reliability of machine-generated insights.

In conclusion, the rapid advancement of AI in science presents both opportunities and challenges. While AI has the potential to revolutionize research and discovery, it also opens the door to academic misconduct and fraudulent practices. As scientists and researchers navigate this new landscape, the need for transparency, accountability, and ethical use of AI becomes paramount to maintaining the integrity of the scientific method.

Click to comment

You must be logged in to post a comment Login

Leave a Reply