The essential link between publishers, librarians and researchers.

 
Recent

The human side of research integrity

Carsten Borchert and Tristan Petit (Image: HZB/M. Setzpfandt)

Carsten Borchert and Tristan Petit discuss navigating challenges in today’s research landscape

The research ecosystem is in constant flux. Maintaining integrity and trust remains at the forefront of advancing knowledge.

While research integrity encompasses many dimensions, such as supervision, data management, and financial stewardship, this article dives into the human side of research integrity. Let’s explore the challenges faced, from honest mistakes in publications to the complexities of collaboration and the importance of transparency in scholarly writing, publishing, and communication.

The psychological burden of honest mistakes

Research integrity is more than simply adhering to ethical guidelines and good scientific practices; it encompasses understanding the psychological and human factors influencing researchers’ behaviour.

One of the most significant challenges authors face is dealing with honest mistakes in their work. Martin Kovacs et al. found that more than three-quarters (82%) of respondents to their survey experienced high emotional and professional stress when their papers were retracted due to unintentional errors1. Of 97 author respondents, nearly half (47%) experienced extreme stress following a retraction, primarily due to the negative perception of retractions from the researcher community. 76% of respondents adjusted their research workflows after experiencing a retraction.

With lapses in attention, technical problems, or communication breakdowns, the psychological impact of these mistakes underpins the need for a more supportive research culture. We need to acknowledge the human side of inquiry and distinguish between misconduct and errors that are honest and not deliberate.

Trust, collaboration, and transparency in research

The research community must collaborate to build on existing knowledge, and trust is essential, but must not be compromised by unreliable studies. Maintaining integrity in research practices is paramount. In collaborative research environments, we recommend:

  • Clearly defining roles and responsibilities is crucial.
  • Effectively interact to foster trust among collaborators.
  • Use a central platform for writing and collaboration to write, comment, and track contributions transparently, keeping the research writing process organised and auditable.
  • Deposit, share and analyse research data openly for data transparently and reproducibly, complementing writing platforms and promoting open research practices.
  • Using efficient communication and data sharing tools to help maintain transparency and facilitate the exchange of ideas and information.
  • Establish clear protocols for data sharing, confidentiality, and intellectual property rights, which are essential for a successful collaboration.
  • Explicitly acknowledge the contribution of each collaborator.

Transparency in research is not just an ethical obligation; it is a cornerstone of research progress. The open research movement has gained significant traction, with more institutions adopting data-sharing policies to promote open scholarship and transparency. This continued shift towards openness redefines research impact by focusing on research value and reuse, improving access and reporting, and enhancing equity in research. Open data sharing is a critical component of open research, as it enhances the reproducibility of studies, exposes data to interrogation for further testing, and allows for the exploration of new research avenues. The open sharing of negative results, which may not be easy to publish through traditional ways, will also facilitate the design of future research. The infrastructure provided by open research enables the sharing of research without limitations, such as space, word count, or paywalls.

Navigate research integrity challenges: from plagiarism to made-up DOIs

Plagiarism, presenting others’ work or ideas as one’s own without proper attribution, remains a serious breach of academic ethics. While many cases stem from poor citation practices rather than intentional fraud, the consequences can be significant, ranging from publication retraction to academic sanctions.

With the rise of AI language models, a new challenge has emerged—machine-generated misinformation. While tools like ChatGPT and other large language models (LLMs) can support writing and research summarisation, they are also prone to generating “hallucinations”—false facts presented with high confidence. The development of AI-generated images also makes it easier to produce and harder to detect fraudulent results2.

A particularly troubling example is the generation of Digital Object Identifiers (DOIs) that do not resolve because they don’t exist3. These alphanumeric strings are designed to provide persistent links to published work, but when fabricated by AI systems, they can give the illusion of legitimate references. This misleads readers and can erode trust in the academic record if such citations go undetected.

The development of AI challenges the classical peer-review process, currently performed by researchers who evaluate the research content of the manuscripts. They must be supported by AI tools to identify fraudulent results, images, and citations that may not be easy for humans to identify. 

The popularity of research misconduct

Research misconduct, including fabrication, falsification, and plagiarism (FFP), remains a continual issue in the research community. Tools like PubPeer, founded initially as an online journal club for authors to discuss published research, have been adopted for specifically discussing and analysing research misconduct in the community.

Microbiologist Sholto David, an active author of For Better Science, has analysed over 1,540 articles using the naked eye and AI tool ImageTwin4. In March 2024, his actions resulted in four retracted articles from Columbia University cancer surgeons5.

“How many errors are acceptable before we think something more worrying is happening?…If you comb through a lot of people’s papers you will find errors, but there has to be, at some point, a limit to how many sloppy mistakes you make before it’s something else, that it’s not something you can dismiss as an honest error” […] “In many cases, it’s simple error rather than fraud…But there’s no bright dividing line. Science isn’t about being perfect. But then you’ve got someone with loads of papers filled with simple mistakes, and you think that’s crossed the line into utter carelessness.” These community initiatives, actions, and continued findings underscore the importance of vigilance and proactive measures to maintain research integrity across the research ecosystem. Every scientist can play an active role in reducing honest mistakes, so that intentional research misconduct becomes more visible. Good practices with AI tools must be defined by experienced scientists, so that improper use of these tools does not become standard habits. On the other hand, self-retraction or article correction must become standard habits, so that honest mistakes can be more easily corrected.

Technological solutions for integrity

Nowadays, technology plays a pivotal role in maintaining research integrity. Platforms and tools have been developed across the research ecosystem from service providers, societies, member bodies, not-for-profits, community initiatives and publishers to assist researchers in ensuring that their work adheres to ethical standards and is free from misconduct:

  • Reference managers

Tools like Zotero, Mendeley, and EndNote help researchers manage citations accurately and consistently, reducing the risk of plagiarism and improper attribution.

  • Electonic Lab Notebooks (ELNs), open data platforms, and repositories

Platforms such as Zenodo, NOMAD, Open Science Framework (OSF), and Jupyter Notebooks facilitate sharing datasets, experimental protocols, and even negative results, encouraging openness, transparency, and reproducibility.

  • Writing quality and citation tools

Language assistants like Grammarly and Writefull help ensure clarity and professionalism in writing. Integrated citation suggestions promote proper referencing and support good scholarly practice.

Grounded.ai helps ensure scientific accuracy and proper sourcing by automatically verifying claims in a manuscript and linking them to credible references.

SciFlow ensures transparency, structured writing, authorship tracking, and reproducible formats like JATS-XML, as well as the integration of the mentioned quality tools.

Plagiarism and image manipulation detection and prevention

Publishers and institutions widely use tools such as Crossref Similarity Check, iThenticate, and Drillbit to detect plagiarism and flag potential issues. Specialised tools for image detection are also available (e.g., ImageTwin and Proofig AI). While no tool is perfect, automated checks support better practices and editorial oversight.

Best practices for maintaining research integrity

To navigate the challenges of research integrity, we recommend these best practices to the research community:

  1. Clearly define roles and responsibilities: establish a comprehensive research plan that outlines methodology, data collection, and analysis procedures.
  2. Communicate effectively: implement regular meetings and use practical communication tools to maintain transparency and facilitate the exchange of ideas.
  3. Give credit, where credit is due: agree on authorship criteria from the project’s outset to avoid disputes later on, e.g., who will be the corresponding author, who will be the first author listed on the paper, and who else will be involved. The CRediT (Contributor Roles Taxonomy) allows authors to provide a detailed and standardised description of their contributions to a research project by distinguishing roles such as Conceptualisation, Data Curation, Writing – Original Draft, Methodology, and more.
  4. Manage and share data to foster open research: implement clear guidelines for data sharing and confidentiality, create a data accessibility statement, and draft collaboration agreements to safeguard the interests of all involved.
  5. Promote diversity and inclusivity: interweave inclusive practices into everything you do rather than making them an afterthought. Encourage diversity within research teams to foster fresh perspectives and innovative ideas. Diversity is more than just cultural heritage and gender. It’s about someone’s experiences in work and life, their beliefs, their different abilities, and the skills they can bring.
  6. Embrace interdisciplinary approaches: address complex challenges from multiple angles, with different voices to provide better context and more informed decisions.
  7. Undertake ethical training: adhere to ethical guidelines and undergo training in responsible conduct of research (RCR) to foster a culture of integrity. Most research institutions and publishers offer support in this area and have expectations around ethical practices for their journals. 
  8. Implement a conflict resolution mechanism: create a structured approach to addressing conflicts, such as designating a mediator or establishing a conflict resolution protocol.
  9. Be transparent and accountable: foster a culture of research integrity that incorporates training, benchmarking, and incentives for continuing assessments and quality improvements.
  10. Use available tools on the author side: leverage technology to embed integrity into your workflow:
  1. SciFlow: Ensures transparency, structured writing, authorship tracking, and reproducible formats like JATS-XML.
  2. Grounded.ai: Provides real-time fact-checking and links claims to credible sources.
  3. iThenticate/Drillbit/Crossref Similarity Check: Helps detect plagiarism.
  4. Zotero, Mendeley, EndNote: Support accurate citation practices.

These tools do not replace good judgment but make building an ethical, transparent, and accountable research process easier. 

As we navigate the complex landscape of research integrity, it is crucial to recognise how human elements influence research conduct. From honest mistakes to the challenges of collaboration and the importance of transparency, researchers face numerous obstacles to maintaining high standards of integrity. By embracing technological solutions, adhering to best practices, and fostering a culture of openness and support, we can create an environment where research integrity succeeds. 

As the founders of SciFlow, Frederik and I (Carsten) developed the platform to simplify the research writing process and support researchers like Tristan. Drawing from our research experiences, we recognise the critical importance of enhancing scholarly workflows to serve the academic community better. We are committed to providing innovative tools and platforms that promote and uphold research integrity. 

As we move forward, we must remember that research integrity extends beyond compliance, regulations, and tools. It reflects a collective commitment to truth, transparency, and the advancement of society through scientific discovery and the sustained support of those who pursue it.

Carsten Borchert is the CEO and co-founder of SciFlow; Tristan Petit is a research group leader at Helmholtz-Zentrum Berlin für Materialien und Energy

References:

1. https://doi.org/10.1098/rsos.240844

2. https://www.nature.com/articles/d41586-024-03542-8

3. https://scholarlykitchen.sspnet.org/2025/03/11/guest-post-trying-to-write-a-paper-with-llm-assistance/

4. https://doi.org/10.1101/2023.09.03.556099

5. https://www.nytimes.com/2024/03/20/science/columbia-cancer-research-retractions.html

Do you want to read more content like this? SUBSCRIBE to the Research Information Newsline!

Back to top