Research Integrity in the Age of AI: Challenges and Guidance for Faculty
As new tools reshape scholarship, Northwestern is strengthening guidance and support for faculty to protect the integrity of the scientific record.
At Northwestern, efforts to improve research integrity acknowledge that AI‑related risks are real and growing but also recognize that they are part of a larger set of pressures transforming the scholarly ecosystem.
In February 2024, a peer‑reviewed biology article was withdrawn after readers noticed AI‑generated figures filled with nonsensical labels and incorrect diagrams. The manuscript looked polished and authoritative until closer inspection revealed fundamental flaws. The swift retraction sent a clear signal: generative AI can now produce research that on the surface looks credible and can be mistakenly published without detection.
A System Under Strain
AI is colliding with an academic publication system already operating at high speed. Scientific retractions have risen sharply over the past decade, surpassing 10,000 worldwide in 2023, driven by a mix of honest errors, data problems, compromised peer review, and potential research misconduct. This reflects both increased vigilance and significant pressure to publish both rapidly and frequently.
At the same time, public attitudes toward expertise are shifting. Scholars have noted a growing resistance to professional knowledge—fueled by online platforms that often reward confidence over correctness, volume over verification. (Tom Nichols calls this the “death of expertise.”) In this environment, qualified statements and methodological nuance struggle to compete with definitive claims and viral commentary.
Surveys show that public support for federally funded research remains strong, even as skepticism about bias, credibility, and academic agendas has increased. The result is a paradox: research is more essential than ever, yet more readily questioned.
Why Integrity Matters
Research integrity—which includes honest methods, accurate reporting, and responsible conduct—is what distinguishes scholarship from speculation. Northwestern defines research misconduct as fabrication, falsification, plagiarism, or serious departures from accepted practices, with consequences that can include retractions, sponsor actions, repayment of funds, and reputational harm.
As a recipient of sponsored funding, Northwestern has a responsibility to ensure research is conducted ethically and in compliance with federal and sponsor requirements. As AI accelerates both discovery and risk, the University continues to reinforce guidance, training, and oversight to support rigorous, reliable science.
AI as Risk—and Safeguard
AI is neither hero nor villain. AI can generate convincingly fabricated content, but it can also be useful for detecting image manipulation, text reuse, and data irregularities.
Northwestern researchers have access to useful AI tools and to guidance on how to use them responsibly. However, it is important to remember that issues flagged by these AI tools are not findings. They signal areas for review, not conclusions. Human expertise, thorough analysis, and context remain essential.
“AI is a powerful tool, but it needs to be used responsibly and with transparency,” said Lauran Qualkenbush, Senior Director and Northwestern’s Research Integrity Officer. “The role of our office is to ensure that concerns are evaluated fairly and consistently with the institutional and federal policies. If faculty face concerns about their research, they should contact our team. We’re here to help navigate the process and protect both the science and the researcher.”
Public Scrutiny and Due Process
Post‑publication review increasingly plays out in public forums. Online scrutiny can surface real issues, but it can also amplify misunderstandings before journals or institutions complete formal reviews. For some researchers, responding to online criticism has become a major time burden, diverting attention from research, teaching, and mentoring.
In a climate already skeptical of expertise, institutional support is critical—protecting due process, safeguarding trainees, and ensuring concerns are assessed based on evidence rather than online momentum.
Sustaining Trust
Paper mills, predatory journals, political polarization, and misinformation campaigns predate AI, but emerging technologies have accelerated their impact. Protecting the research requires more than compliance; it requires clear lab practices and data management, transparency in performing and reporting the research, and shared responsibility.
“AI offers meaningful opportunities to advance discovery, yet it also calls for careful use, supported by strong guidance and a culture that values thoughtful, responsible research practices,” said Vice President for Research Eric Perreault. Staying informed, asking questions early, and engaging with the Office for Research Integrity when uncertainty arises are essential to sustaining trust in Northwestern’s research. — Matt Golosinski