Early May, I attended the 2026 World Conference on Research Integrity (WCRI) in Vancouver, BC, Canada. This conference is held every two years. In 2024, I went to the WCRI in Athens, Greece, and in 2022 to the one in Cape Town, South Africa. In this post, I will provide a short summary and links to my BlueSky #WCRI2026 posts for all the sessions I attended.
The WCRI 2026 was held in the Westin Bayshore, next to Vancouver Harbor. This hotel has a decent-sized conference center, and could easily fit the ~700 attendees for this event.
Sunday May 3rd, Pre-conference Workshops
The conference had 7 pre-conference workshops on Sunday, of which I attended two.
On Sunday morning, I attended Workshop 2 — AI in Bioimaging: Challenges and Solutions to Facilitate FAIR Image Data. Speakers Jana Christopher and Nathalie Gaudreault discussed the growing challenges around image integrity, metadata, and AI-generated scientific images. A major theme was that scientific images should be treated as research data, not just illustrations. The workshop highlighted the importance of metadata, raw data preservation, reporting standards, and new tools such as BioFile Finder to improve reproducibility and trust in bioimaging research.
Here are my BlueSky posts:
After lunch, I went to Workshop 7 — Publication Integrity & Institutional Investigations. Speakers Lisa Leventhal, Daniel Ucko, Daniel Barr, Mai Har Sham, and Tom Lindemann discussed the often difficult interactions between publishers, journals, and institutions during misconduct investigations. Key topics included slow investigations, lack of research integrity infrastructure, confidentiality barriers, inconsistent institutional responses, and the need for better international coordination and communication between publishers and research institutions.
Sunday May 3rd, Conference Opening Ceremony
The WCRI 2026 opening ceremony began with a welcome honoring First Nations culture, including a traditional song and an Indigenous Hoop Dance performance by Jay (I did not catch his last name). Conference co-chairs David Moher and Chris Graf welcomed more than 700 attendees from 57 countries and highlighted major conference themes including artificial intelligence, research security, and Indigenous knowledge systems.
Speakers Mona Nemer (Chief Science Advisor of Canada) and Alejandro Adem (NSERC) emphasized the importance of public trust, science integrity policies, transparency, mentorship, and responsible use of GenAI and Open Science. Ry Moran spoke about reconciliation, Indigenous participation in research, and the importance of including diverse voices in science. The evening concluded with the IEEE-sponsored welcome reception.
Monday, May 4th, morning sessions
Plenary Session 1 — Misuse of AI by Paper Mills
Speakers Daniel Moreira, Stefan Stender, and Renee Hoch discussed how generative AI is accelerating paper-mill activity and enabling large-scale production of fraudulent or low-quality scientific papers. Topics included AI-generated scientific images, synthetic Western blots, “fast churn” database studies, and the mass production of Mendelian randomization papers using open datasets and automated tools.
The speakers emphasized that AI is making fraudulent content both easier to produce and harder to detect, creating an “arms race” between generation and detection. Possible solutions included image-forensics tools, stronger peer review, better screening systems, primary-data requirements, and shifting incentives away from publication counts and paper metrics.
Parallel Morning Session — Questionable Journals, Paper Mills, Law & Research Security
After the plenary morning talks, there were several parallel sessions to attend. It was hard to make a decision, because talks were not really grouped per topic, so most sessions covered several unrelated but interesting topics. I just picked one that I thought had the most interesting talks.
Daniel Acuna presented AI and bibliometric approaches to identifying questionable journals, using citation patterns, DOAJ inclusion, and publisher behavior as signals. Anastasiia Balkina discussed paper mills infiltrating IEEE conference proceedings, including patterns such as six-author papers, multiple affiliations, and “bio-inspired” algorithm papers. Kritika Sharma spoke about similarities and differences between legal investigations and research misconduct investigations. Lucas Tersigni presented Canadian research security trends, highlighting risks in AI, quantum science, and green engineering partnerships.
Monday, May 4th, afternoon sessions
I hopped a bit around the different rooms.
Sabina Alam from Taylor & Francis discussed growing integrity threats faced by publishers, including image and data manipulation, fake reviewer identities, bribery attempts targeting editors, and fake journal websites. She emphasized that confusion remains widespread about what third-party editing or publication services are acceptable, especially regarding paper mills and AI-assisted writing. The session also highlighted collaborative integrity initiatives between publishers, institutions, and libraries, including recommendations for clearer integrity training and responsibility-sharing. A recurring theme was that many researchers still lack formal integrity education.
Rosemarie De La Cruz Bernabe presented a critical perspective on Open Science, arguing that “openness” is not automatically neutral or equitable. Concerns included APC inequality, infrastructure gaps, invisible labor, and “data colonialism,” where benefits of openness are distributed unevenly.
René Aquarius presented a large-scale analysis of preclinical osteosarcoma papers using ImageTwin and PubPeer screening. Out of more than 1,000 papers examined, about 12% showed problematic image duplication or manipulation. Certain journals appeared disproportionately affected, with Oncotarget described as likely having been targeted by paper mills.
The talk stressed that automated tools were only a first-pass screen and that all suspected cases were manually checked. The analysis suggested strong clustering of problematic papers within certain journals and countries.
Another talk explored the environmental impact of unethical AI practices in research. Suresh Baral modeled the energy and carbon costs associated with large-scale AI misuse and fraudulent computational work, arguing that integrity failures can also create measurable environmental waste.
After the tea break, I chaired Oral Session 3D, which covered topics such as AI tools for detecting duplicate peer reviews, large-scale mining of Crossref metadata for integrity concerns, AI-generated academic image manipulation, analyses of retracted RCTs, and understanding of plagiarism among international trainees.
Because I was moderating this session (which is hard work with 10-min talks!) I did not have time to post much on BlueSky, but here are my posts.
Monday, May 4th, Evening Steneck-Mayer Lecture
The evening Steneck–Mayer lecture was delivered by Robbert Dijkgraaf on “Science under pressure in a fractured world.” He argued that science is simultaneously flourishing technologically while facing increasing political, economic, and societal pressures, including censorship, distrust, intimidation, and AI-enabled misconduct.
Dijkgraaf called for reform of incentive systems, stronger peer review and replication efforts, early education in integrity, clearer AI disclosure policies, and active defense of international collaboration and academic openness.
I very much enjoyed this lecture – Dijkgraaf is an excellent and motivating speaker.
Tuesday, May 5th, Morning
The morning plenary focused on “GenAI and Research Integrity: Global Ethics, African Wisdom, Shared Responsibility.”
Lorna Waddington presented a nuanced GenAI talk, emphasizing both usefulness and unreliability. She showed examples where LLMs pulled misinformation from extremist websites and argued that AI systems inherit and amplify biases already embedded in their training data. Examples included LLM refusals to process historical Nazi-era documents, and image-generation oddities such as Gemini refusing “German soldier 1940” prompts while allowing nearby years.
Titilola Olojede introduced the Yoruba concept of “Ọmọlúwàbí” as an integrity framework rooted in communal responsibility and character. Rather than asking “Can I get away with this?” the framework asks “What kind of researcher does this make me?” and “Who is harmed by my silence?”
Retha Visagie argued that conventional research integrity systems are heavily shaped by Western traditions and may exclude indigenous or community-centered knowledge systems. She emphasized reciprocity, solidarity, and the need for AI governance frameworks adapted to African realities, noting stronger reliance on GenAI tools in parts of the Global South.
After the coffee break, I attended oral session 4D. Nancy Kwangwa presented interviews showing that many researchers do not disclose AI usage and often see research integrity mainly as plagiarism avoidance rather than an everyday research practice. She proposed a practical AI disclosure template.
Svetlana Kleiner discussed growing integrity concerns in edited books and monographs, arguing that books may become “the new special issues.” Cases included suspected compromised peer review, undisclosed LLM use, and tortured phrases. She noted that publishers often cannot definitively prove LLM use, making retractions difficult, and stressed that “vibes alone are not enough.” Springer Nature recently introduced Expressions of Concern for books.
Joonha Jeon discussed paper mills as organized systems rather than isolated misconduct cases, highlighting markers such as duplicated images, suspicious textual patterns, reused datasets like NHANES, and coordinated citation behavior.
Sarah Jenkins described Elsevier’s efforts to detect large-scale networks manipulating citations, editorial roles, and peer review. Elsevier reportedly investigated 24 bad-actor networks involving about 1,100 members across journals. Audience members questioned why such problems were detected relatively late despite Elsevier’s large resources.
The session concluded with talks on doctoral defense misconduct in Mongolia by Orkhon Gantogtokh (UBC) and research misconduct surveys in Ethiopia by Habtamu Hailu (Ethiopian Defence University), including concerns about favoritism, publication pressure, and high self-reported misconduct rates.
Tuesday, May 5th, Afternoon
I was part of Plenary Session #3 — Einstein Foundation Award Winners. This afternoon plenary featured presentations by winners of the Einstein Foundation Award for Promoting Quality in Research.
After an introduction video talking about the scope of this award, Timothy Errington discussed large-scale reproducibility studies in the social and behavioral sciences, including repeatability analyses of hundreds of papers across dozens of journals.
Brandon Stell presented PubPeer as a global, community-driven post-publication review system that functions as a large online journal club where researchers can share concerns and evaluations of published papers.
Finally, I (Elisabeth Bik) discussed how image-integrity investigations revealed widespread problems with duplicated images, paper mills, and AI-generated fake data. I argued that scientific misconduct is increasingly rewarded, while corrections remain slow and whistleblowers often face harassment and legal risks.
Next, I attended the symposium Sleuths and Publishers Collaborating to Preserve Research Integrity. This symposium focused on collaboration between publishers and research-integrity sleuths. Chair Adya Misra emphasized that paper mills and GenAI are accelerating misconduct while trust in science is becoming more fragile.
René Aquarius discussed the slow pace of post-publication correction processes, noting that many problematic papers remain uncorrected for years despite clear image duplication concerns. He also described increasing pessimism about the scale and sophistication of paper mills.
Laura Wilson outlined practical lessons from publisher investigations, including the importance of polite communication, confidentiality, right-to-reply, and “Under Investigation” banners. She also described challenges including AI-generated complaints, aggressive correspondence, overwhelmed editors, and lack of institutional responses.
Adya Misra argued for stronger collective action between publishers, institutions, and sleuths, including better engagement with sleuth-generated evidence and more scalable integrity workflows.
Finally, Achal Agrawal (online) proposed a private “Sleuth Box” platform for confidential communication between sleuths and publishers, while acknowledging that such a system could not replace the public transparency provided by PubPeer.
Tuesday, May 5th, Evening Dinner at the Vancouver Aquarium
On the evening of May 5th, the WCRI conference dinner took place at the Vancouver Aquarium in Stanley Park. The aquarium, founded in 1956, is one of Canada’s best-known marine science and conservation centers, featuring Pacific Northwest marine life, tropical fish, jellyfish, sea otters, and seals.
The dinner required a separate ticket, which several attendees felt was relatively expensive, leading some sleuths and colleagues to organize alternative evening plans instead. Still, the venue itself was widely appreciated, and the evening became a relaxed opportunity for informal networking and conversations outside the conference setting.
Wednesday, May 6th, Morning
The morning session focused heavily on paper mills and integrity screening tools. Gengyan Tang presented research on paper mills in China, describing how ghostwriters, publication pressure, and broader structural incentives sustain the industry. He argued that addressing paper mills requires systemic change, including reducing publication pressure on clinicians.
Colby Vorland introduced INSPECT-AI, a semi-automated tool to help reviewers identify integrity concerns in randomized controlled trials, while keeping human judgment central.
Later talks included Amanda Sulicz describing IEEE’s investigation into a suspected paper-mill-linked conference, and Michael Streeter discussing Wiley’s integrity screening workflows for misconduct, peer review manipulation, and ethics concerns.
Wednesday, May 6th, Afternoon
The afternoon session covered Open Science, AI in research, and research ethics in different cultural contexts. Inge Stegeman discussed the OSIRIS project studying whether Open Science practices actually improve reproducibility.
Mads Sørensen presented survey results showing that attitudes toward GenAI vary strongly across disciplines, arguing against one-size-fits-all guidance.
Nihar Shah gave a spectacular presentation that started with a completely dark room, the sound of bells, and the speaker appearing in a Grim Reaper costume. He warned that autonomous “AI scientist” systems can introduce integrity problems including benchmark selection bias, data leakage, and automated p-hacking behavior.
Additional talks addressed ethical challenges in adolescent health studies in Bangladesh, AI-assisted forensic analysis, and conducting research ethically in rural South African communities.
After the final afternoon sessions, the conference moved into the closing ceremony, opened by conference chair Chris Graf. He thanked the many organizers, reviewers, sponsors, technical staff, and volunteers who helped run the meeting, and humorously taught the audience a “bodily drumroll” before the award announcements.
A large number of awards were presented, including postdoctoral, oral presentation, poster, and diversity awards. Among the Anderson-Kleinert Diversity Award recipients were Allen Mukhawana, Vladyslava Kachkovska, and Lisa Dixon.
Zoë Hammatt thanked several retiring board members, including Sabine Kleinert, Lex Bouter, and Lynn, recognizing their foundational contributions to the research integrity field. Each received a locally made wooden artwork.
One of the highlights of the closing ceremony was the announcement that the next World Conference on Research Integrity will take place in [body drumroll, please]…. Tallinn, Estonia, in 2028. Margit Sutrup welcomed attendees to the future conference, joking that she could not promise beautiful weather but could promise a wonderful city. She also shared a personal connection between Vancouver and Tallinn through her family history.
The ceremony concluded with final thanks to the sponsors, organizers, and participants, followed by a group photo of the many award winners.
I hope you all enjoyed following my live posts from WCRI 2026 on Bluesky. It was a fascinating and sometimes sobering conference, with many important discussions about AI, paper mills, research culture, reproducibility, and the future of science integrity. It was great that the conference offered a podium for both sleuths as well as investigations by publishers into paper mills and other integrity problems. In previous years, the conference sometimes was a bit self-congratulatory (look how great we publishers/institutions are doing). This year, the themes were more realistic – and frightening.
I also enjoyed having drinks and dinners with some old and new friends. A special mention should be made of the Red Accordion, which became the official #WCRI2026 watering hole of the ‘sleuths’.
Thank you to everybody who followed along, replied, shared information, or came up to say hello during the meeting — and hopefully I will see many of you again at WCRI 2028 in Tallinn, Estonia.
Science Integrity Digest 