Million-Dollar Project Aims to Expose Bad Medical Research

A new initiative of Watchdogs on the back of the withdrawal clock is targeting a defective or fake medical science research of nearly $ 1 million.

The The center of scientific integrity The only medical proof project has launched the project, two years of effort to detect treatment research with negative impact on health guidelines-and to ensure that people have actually heard about it.

The project is equipped with $ 900,000 grants and five investigators from Open Philantropi, the project will use forensic metasians tools to identify scientific article problems and report its results through the maximum site withdrawal watch for scientific surveillance.

“We have originally set up the center of scientific integrity as a house for the withdrawal watch, but we always hope that we will be able to do more in research accountability,” the Center’s Executive Director and Co-Clock co-founder Ivan Onsky said, Post Declare grants. “Medical Proof Project allows us to support criticism analysis and promote the search.”

Accordingly NatureThese defective and false documents are wandering because they have meta-analyzes skio-reviews that combine multiple research results into a more statistically strong decision. If the study of one or two banks turns it into a meta-analysis, they can press the fibers in the health policy.

In the 21st, to name a case, a European guideline suggested the use of beta-blockers during non-Cardiac surgery, based on the Millennium research that was later called in question. A few years later, an independent review has suggested that the guidance can contribute to 10,000 deaths every year in the UK.

The team’s plan led by the integrity adviser to the science is to create software equipment, chase the leadership from the whistle blowers anonymously, and pay the peer reviewers to test their work. They aim to detect at least 10 defective meta-analyzes a year.

The team is choosing his moment with wisely. Such as Gizmodo had previously reported, AI-exposed junk science The academic digital ecosystem is flooding by showing everything from the conference activities to the Pierre-Paalo Journal. A study published in the Harvard Kennedy School’s misinformation review that two-thirds of the sample papers obtained through Google Scholars contain signs of GPT-exposed text-some even in mainstream scientific outlets. About 14.5% of this bogus studies focus on health.

This is especially worrying because Google scholars do not distinguish between peer-polluted studies and prints, students’ papers or other less-tait. And once this type of bike is drawn in meta-analysis or when cited by physicians, it is hard to endure the consequences. A researcher told Gizmodo, “If we can’t believe that the study we have read is real,” we risk making decisions based on misinformation. “

We have already seen how the bad thing can be left behind. In the 2021, the nature of the Springgar More than 40 withdrawal The papers of its Arabian Journal of Geoscience-Studies were so unscrupulous they read like AI-exposed Mad Libs. Just last year, the publisher borders had to pull a paper that featured a physically impossible AI-exposed images Rat genitaliaThe

We have entered the era of digital fossil, where the AI ​​models trained on web-scraped data have begun Preserve and promote Bad phrases were as if they were true scientific terms. For example, earlier this year a group of researchers found a garland set of 1959 biology paper that were embedded on the outputs of large language models, including OpenAE’s GPT -4O.

In that climate, the goal of the medical proof project is felt more than a trize than cleanup. The team is one of the defective information, hiding in simple sight and having a lot of health consequences if the face price is taken, it can be a very real consequence.