The process of getting a scientific paper peer reviewed isn’t exactly easy. In order for genuine progress to be made in any given field, the research upon which that progress is based has to be thoroughly tested and reviewed.
Often, this peer reviewing process is a labor of love that goes unpaid. With how many academic papers get published, getting enough people to peer-review them was already a struggle before the advent of AI slop.
But, now that several publishing platforms have been flooded with literal millions of papers at once, people who are actively trying to vet these papers and review them are hitting an impossible wall.
There’s no way to feasibly sort through the sheer quantity of slop hitting publishing platforms for scientific papers, and researchers are in a panic.
AI slop papers take over research publishing platforms
A paper authored by several authors in late 2024 titled “The strain on scientific publishing” outlined some of the crippling and possibly unsolvable issues in the field of scientific research.
“Scientists are increasingly overwhelmed by the volume of articles being published. The total number of articles indexed in Scopus and Web of Science has grown exponentially in recent years; in 2022, the total was ∼47% higher than in 2016, which has outpaced the limited growth—if any—in the number of practicing scientists. Thus, publication workload per scientist has increased dramatically,” reads the paper’s abstract.
And, while there are a number of factors contributing to this trend other than the generation of AI articles, it certainly isn’t helping.
Some papers are getting put up for peer review that are purely AI-generated, with seemingly no one checking the paper for quality control before sending it into the ether to get peer reviewed.
The issue came to a head when a scientific paper published an AI-generated study involving rat genitals, creating a truly nightmarish image that’s a bit hard to describe.
We’re using this picture of a low-poly rat from Halo 3 to spare you from having to see AI-generated horrors
But, even in the year and a half since that paper got pulled off the journal it was published in, AI has gotten a lot better at creating images that are a bit more believable.
“Everybody agrees that the system is kind of broken and unsustainable,” veteran scientist Venki Ramakrishnan told The Guardian. “Nobody really knows what to do about it.”
With literal millions of papers getting published that may never have the time to get a pair of human eyes on it, the ability to authentically and organically sort through the amount of research papers getting published is getting further and further away from feasibility.
Source link