AI scandal rocks academia as nearly 200 studies appear to have been generated in part by ChatGPT
The academia was engulfed in scandal this week after it emerged that research articles in dozens of scientific journals were written by AI.
An investigation found more than 100 articles likely written, at least in part, by ChatGPT.
These articles have slipped through the cracks due to lax or non-existent peer-review processes at for-profit journals, fueling fears that human scientific knowledge is rapidly being infiltrated by low-quality computer-generated garbage.
Many of these articles are published in obscure scientific journals, but news of this type of scientific fraud damages public trust in science, many scientists agree. Even before AI-generated text appeared in scientific articles, so-called “paper mills” have been churning out low-quality and even plagiarized articles for years.
This article, published in the International Journal of New Media Studies, bears the telltale sign of being written by AI: “As of my last knowledge update.”
This chapter from a book on media studies addresses issues important to youth in India. It contains strange sayings, such as ‘Youth is the time of life when you are young.’
The influence of AI on research papers was exposed by a new report from the tech journalism site 404 Media.
115 results were found for ‘From my last knowledge update’ in Google Scholar, Google’s search tool for academic articles. This phrase, which returned 188 results as of Tuesday, is often used by the bot to tell users how current the information is.
404 Media reported that the dates in the newspapers containing the phrase matched the dates of actual ChatGPT knowledge updates.
The articles covered topics such as spinal injuries, battery technologies, rural medicine, bacterial infections, cryptocurrency, child welfare, and even artificial intelligence.
Many of the journals that publish these fraudulent articles rake in money and charge authors processing fees to publish their articles that contain no new ideas.
A few of the search results are not intended to fool anyone, and are in fact about the pitfalls of using ChatGPT as a research tool.
‘Don’t Trust ChatGPT: A Case Study of a Defective Research Tool’ is the title of such an article.
Removing results with ‘ChatGPT’ in them still yields as many as 135 articles that contain a strange phrase.
Some of these articles seem legitimate, while others are barely understandable: ‘Global Education Education and International Education Advocacy’ is one such example of unreadable titles.
The phrase even appears in a paper about how AI systems can be made more transparent and explainable.
Bellingcat researcher Kolina Koltai Posted a screenshot of an academic article about
Besides the fact that using ChatGPT is easy, professional pressure is a major reason this happens.
ChatGPTis is regularly updated with new knowledge about the world. When it answers questions, it often becomes clear when the knowledge was last updated. It’s a useful note for ChatGPT users, but it also provides a clue that an academic article was written by ChatGPT.
Bellingcat researcher Kolina Koltai posted this screenshot on
Academic researchers face enormous pressure from their universities to publish articles because they are one of the main ways scientists are judged when it comes time to get a new job or promotion.
This challenge is so common that it has an ominous expression: “publish or perish.”
And for those who want to take the easy way out, there are plenty of scientific journals more than willing to accommodate.
Normally, the publication process for a renowned academic journal is lengthy, requiring peer review by other scientists, often followed by multiple rounds of revisions.
Scientists may even need to conduct additional experiments to fulfill requests from journal editors and reviewers.
This process can take months or even years, depending on what changes are needed.
This article contained two different versions of the meaningful phrase. The magazine it is published in charges a processing fee of just $4.
This scientific article appeared in December. It has since been cited zero times, suggesting that other researchers did not find it useful
But overall, the end result is a solid piece of work that has been carefully researched, polished and perfected.
That’s not the case with these “paper mills” that accept almost any submission – as long as the author also pays the publishing costs.
For example, the International Journal of New Media Studies published two different articles with the telling phrase: “As of my last knowledge update.”
The journal claims to conduct peer review and even has a page dedicated to describing the process.
“The practice of peer review is to ensure that only good science is published,” reads the introduction to the page on the magazine’s website.
‘It is an objective process that is at the heart of good scientific publishing and is carried out by all renowned scientific journals.’
Peer review involves sending a manuscript to several experts – usually three – who read the article carefully and express their criticism. Even without major revisions, this process takes time, because reviewers are working scientists who must make time in their schedules to read the article.
This article is entitled ‘Global Education Iducation and International Education Advocacy.’ It was published in January this year and contains only four references to other studies
A second use of the keyphrase in this article shows that it was almost certainly written by AI
Still, the journal promises that articles will be published one to two days after submission.
DailyMail.com has contacted the magazine’s editors about its policy on AI-generated submissions.
The International Journal of New Media Studies charges authors $50 to publish an article, another $20 to assign the article a DOI number (used for online indexing), and another $50 to obtain a hard copy of it receive a magazine.
This magazine is a striking example of this, but not unique.
Searching for ‘From my last knowledge update’ will bring up a huge list of articles published in a variety of journals, on topics as diverse as materials science, youth wellbeing and quantum physics.
It’s no secret that paper mills produce low-quality paper for a fee.
But as AI chatbot tools become easier to use and more widely available, these low-value papers only seem to become more plentiful.