IEEE continues to strengthen its research integrity process

Ensuring the integrity of IEEE Research Publishers is critical to maintaining the organization’s credibility as a scientific publisher.

IEEE produces more than 30 percent of the world’s professional literature in the field of electrical engineering, electronics, and computer science. In fact, the 50 companies with the best patents are cited by IEEE nearly three times more than any other technical literature publisher.

As the volume of academic papers submitted has increased over the years, IEEE is constantly evolving its publication processes based on industry best practices to help identify problem papers. These include plagiarism, inappropriate citations, coercion of authors by editors or reviewers, and the use of artificial intelligence to generate content without proper publication.

“There are now many more types of misconduct and many more cases of infringement across the publishing industry, so it’s critical for all technical publishers to take it seriously,” says IEEE Member Sergio Benedetto, vice president of IEEE Publication Services. and Products.

“Authors are also more careful in choosing publishers who are serious about dealing with errors. It has become important not only for the integrity of research, but also for the publishing business,” adds Benedetto, IEEE board member.

“It’s important to understand that IEEE is not blind to this problem, but rather is investing heavily in trying to solve it,” says Steven Heffner, executive director of IEEE Publications. “We are investing in detection and investigation technology. We also invest with partners in the development of their technologies to help the entire industry expand the detection system. And we’re also investing in human resources.”

New ways to set up the system

Some of the main causes of misconduct have to do with incentives offered to authors to encourage them to publish more, Heffner says.

“Promotion and tenure are tied to publishing research, and the old ‘publish or perish’ imperative still applies,” he says. “Now there are more ways scientists can be evaluated, with technology tracking how many times their work has been cited and with bibliometrics, which is a statistical method that counts the number of publications and citations by an author or researcher. These statistics are used to verify the value their work brings to their organization.

“Even more sophisticated ways are used to manipulate the bibliometric system,” says Heffner, “such as pseudo-citation cartels that operate in a quid-pro-quo ‘I’ll cite you if you cite me.’

“Unfortunately, all of this creates more opportunities for people to abuse the system.”

Other activities on the rise include paper mills run by for-profit companies that create fake journal articles that look like real research and then sell the authorship to would-be scientists.

“There are now many more types of misconduct and many more cases of infringement across the publishing industry, so it’s critical for all tech publishers to take them seriously.” —Sergio Benedetto

“I think the paper mill is the most dangerous problem at the scale we have,” Heffner says. “But old crimes like plagiarism still persist and in some ways are becoming harder to detect.”

Benedetto says some fraudulent authors make up the names and websites of reviewers so that their articles are accepted without going through the peer review process. It’s a serious problem, he says.

“I don’t think IEEE is unique in its experience with this phenomenon of misconduct,” he says. “Several commercial publishers and many in non-tech areas are seeing the same issues.”

Resolving author errors

The IEEE PSPB Publishing Content Committee, which deals with cases of editorial misconduct, considers violations of its publishing process to be serious offenses.

“IEEE volunteers are particularly strong in policymaking,” says Heffner. “We need that management, but we also need their expertise as people involved in the scientific endeavor.”

Benedetto says the IEEE is serious about finding questionable documents and approaches and has launched several initiatives.

IEEE checks all content submitted to journals for plagiarism. Systematic real-time data analysis during the publication process helps identify potential wrongdoing. Paper reviewers are required to include recommended references on their review form to track high bibliometrics.

The organization’s peer review platform works to identify potential abuse by reviewers and editors. It tracks biased reviews, conflicts of interest, plagiarism, and monitors reviewer activity to identify patterns that might indicate inappropriate behavior.

The names of authors and editors are checked against a list of banned participants, people who have violated the IEEE’s publishing policy and can no longer publish in the organization’s journals.

Some unscrupulous authors use artificial intelligence to game the system, Heffner says.

“With the advent of generative artificial intelligence, completely fraudulent papers can be produced faster and appear more convincingly authentic,” he says.

This leads to concerns about the validity of the data.

The new policy addresses the use of artificial intelligence by authors and reviewers. Authors who use AI to create text or other work in their articles must clearly identify the sections and provide appropriate references to the AI ​​tool. Reviewers are not allowed to load manuscripts into a large AI-based language model to generate their reviews, nor may they use AI to write them.

Anyone who suspects any misconduct—including inappropriate citations, use of artificial intelligence, and plagiarism—can file a complaint through the IEEE Ethics Hotline. It is available seven days a week, 24 hours a day. The process is managed by an independent third party and the information provided is sent to IEEE confidentially and anonymously upon request.

Types of corrective measures

If an author or reviewer is suspected of misconduct, a case is opened and a detailed analysis is carried out. An independent commission will review the information and, if warranted, begin an investigation. The alleged offender is allowed to respond to the accusation. If the offender is found guilty, several penalties may be imposed.

Depending on the severity of the violation, an escalating system of sanctions is used. Individuals who plagiarize content at a sufficiently serious level are restricted from editing and publishing and their names will be added to the Participant Prohibited List (PPL) database. Persons on the list may not participate in any activities related to IEEE publications, including conferences. They are also removed from any editorial positions they hold.

The IEEE has strengthened its policies for retracting and removing articles. When an article is flagged, the author receives an expression of concern. Unreliable data may be the result of honest error or research misconduct.

IEEE considers retractions to be a method of literature correction. If there are problems with the content, he takes due care and time in the review and withdraws the non-compliant publications when necessary. Retraction notices alert readers to these publications. Retractions are also used to alert readers to instances of redundant disclosure, plagiarism, and failure to disclose competing interests that may have influenced interpretations or recommendations. In the most severe cases, articles are removed.

Retracted articles are not removed from hard copies of the publication or electronic archives, but their withdrawn status and reason for withdrawal are explained.

IEEE’s remedial measures for violations of publication behavior previously focused on authorship restrictions, but now include restrictions on editorial roles such as reviewer, editor, conference organizer, and conference publication officer. Their names are also added to the PPL and their publication through IEEE may be prohibited.

Industry-wide efforts to uncover misconduct

IEEE and other science, technology and medicine (STM) publishers have joined forces to launch pilot programs aimed at detecting concurrent submissions of suspicious content across publishers, Heffner says.

They are working on the development of the STM Integrity Hub, a powerful submission review tool that can flag tactics related to misconduct, including paper mills.

Publishers are also developing their own tools based on artificial intelligence and machine learning to screen peer-reviewed submissions and articles in real time. Some tools have already been implemented.

Benedetto says he is working on a process to share IEEE’s blacklist with other publishers.

“Those found guilty of wrongdoing simply turn to other publishers,” he says. “Each publisher has its own list, but these aren’t shared with others, so it’s very easy for a banned author to change publishers to get around the ban. A shared list of misconduct would prevent those found guilty from publishing in all technical journals for the duration of their sentence.”

“We’re all working together to share information and to share best practices,” says Heffner, “so that we can combat this as a community of publishers who take scholarly records management seriously.”

“Some colleagues or authors think that misbehavior can be a shortcut to building a better career or achieving publication goals more easily,” says Benedetto. “That’s not true. Misbehavior is not a personal matter. It’s a problem that can and does breed distrust of institutions, publishers, and journals.

“IEEE will continue to strengthen its efforts to combat cases of publication misconduct because we believe that research integrity is at the heart of our business. If readers lose confidence in our journals and authors, they will lose confidence in IEEE itself.”