Guidance on use of artificial intelligence in submissions to Qualitative Psychology

This guidance reflects on the use of generative artificial intelligence (AI) in the processes of qualitative research and manuscript development for authors submitting to Qualitative Psychology (QP). It is important to us that readers continue to trust the research that we publish and that there is transparency about the kind of qualitative research that we view as high quality, as ethical, and as having scientific and methodological integrity. The editorial board expects authors to adhere to this guidance when submitting manuscripts to our journal.

The following are central concerns about the use of generative AI methods within the process of qualitative research:

1. Scientific and methodological integrity concerns regarding fabrication: AI has been known to fabricate information, findings, and citations, with the proportions of misinformation appearing to be escalating over time (Metz & Weise, 2025). For instance, a recent study examined eight AI programs on their data extraction of over 300,000 data points from 22 systematic review databases, finding excellence in accuracy within only 12% of variables and at least some unacceptable metrics within 66% of variables (Jansen et al., 2025). Another assessment of fabrications in a large set of large language models found that the best 25 still had error rates that ranged between 3.3 - 10.9%, with some popularly used models not making the list (e.g., Gemini-3-pro had a 13.6% rate; Awadallah & Mendelevitch, 2025). This means that information gathered by AI is not reliable and, when used in scientific studies, needs to be checked for accuracy via human analysis.
2. Scientific and methodological integrity concerns regarding bias and manipulation: AI can be influenced by public stereotypes and prejudices, by the values and agenda of its creators, as well as by the types of training material that it has encountered; it appears that AI may be deliberately adjusted to reflect certain perspectives or to communicate misinformation for political or other ends (Jones, 2025; Thompson et al., 2025).
3. Scientific and methodological integrity concerns regarding lack of morality and reflexivity: Moral, ethical, critical, and reflexive capacities are necessary to perform strong qualitative research (Braun & Clarke, 2022; Levitt et al., 2021; Shaw et al., 2020) but these features are lacking in AI. AI is unable to examine its positionality critically or seek out information that would balance biases in its training material.
4. Scientific integrity and methodological concerns regarding artificial data drift: Because AI can generate information faster than people, AI in the future may be trained dominantly on AI-produced or synthetic data. This issue can lead to models collapsing and to AI outputs increasingly misrepresenting true data distributions (Shumailov et al., 2024). 
5. Ethical concerns regarding inadequate sourcing: There are ethical concerns about AI because the basis for information it produces is rarely fully documented, and authors or artists often do not give consent for their work to be used as training material (Ortutay, 2025).  Inadequate sourcing can make it difficult for authors to adhere to scientific ethics of citation.
6. Ethical concerns regarding appropriation: Work that is submitted into an AI program may be incorporated into the data that feed the AI program without asking the authors’ consent (Andreotta et al., 2022). These programs may store and use the ideas and works of authors and may not support users to know of or cite those contributions in a way that accords with scientific ethics.
7. Environmental concerns: Concerns about the environmental impact of AI are sharply growing (Yu et al., 2024).Research suggests that many of the environmental costs are still not well known, including its impact on not only electricity demanded but the consumption of fresh water and rare metals (Dungo et al., 2025). Qualitative researchers from across the globe have expressed concern about exploitative and extractive impacts of AI and cautioned against its use (Jowsey et al., 2025).
8. Social justice concerns: These environmental concerns appear to be disproportionately impacting low-income neighborhoods and communities that are on rural or indigenous lands. A recent review of the research on the impact of AI stresses the necessity of not only considering its efficiency gains, but also “hidden or temporally deferred costs that accumulate’ and—notably—the distributional consequences that determine who reaps the benefits and who inherits the burdens” (Ojong, 2025, p. 8305).

For these ethical, environmental, and research integrity reasons, QP encourages authors to avoid using AI in developing scientific articles. The preceding concerns make clear why the fidelity of AI outputs appears questionable at this point. They require caution about the capacity of AI programs to reflect human experience and social practices in an ethical, socially just, and environmentally just manner. 

In keeping with APA policy, QP prohibits reviewers from entering any part of a manuscript into AI or from using AI in the making of any part of the guidance or recommendations that they provide when reviewing a manuscript. We caution researchers to take care that confidential information (e.g., from clinical or research contexts), or data owned by someone else, is not shared with AI programs that store or appropriate their data. Before entering information into AI programs, researchers should know whether the programs are compliant with privacy acts and confidentiality laws in their countries (e.g., HIPAA). Data residency requirements (where data is stored), the availability of audit logs, the level of reproducibility in findings, and evidence on the proportion of fabrications may be beneficial information for researchers to learn as well (Silber, 2026).

We encourage researchers to consider the costs of using AI. Its use can limit the engagement of investigators, which makes qualitative research insightful, ethical, responsive to the complexity of human experiences, and rewarding. In response to environmental concerns, QP editors encourage researchers to seek options for word selection or sentence-level editing that have lower environmental costs and have fewer ethical or data integrity concerns (e.g., using the Thesaurus option in Word or a less environmentally impactful AI program). 

To be specific, QP does not publish manuscripts in which AI was used:

• To conduct literature reviews
• To draft text describing research design and methods
• To collect and/or analyze data or other information
• To draft descriptions of analyses and findings
• To interpret analyses or connect findings to the literature
• As a sole, primary, and/or as an unchecked source for identifying references and citations

If AI was used in a manuscript, a detailed statement making clear that it was not used for these activities listed above and specifying how it was used is required. 

QP is aware of the advantages that AI can have for authors who are non-English speakers and who may face difficulty in publishing in English-language journals. If AI is used in translation, please append a copy of the manuscript in the original language in the online supplementary material. Once the paper is accepted, please include a resubmitted version with a description of changes that were made in the review process and instructions for readers to cite the published version of the paper. The availability of this document can increase the accessibility of that work, especially in the country where the research was conducted. Also, in line with QP’s expectations for such use of AI for translation, please indicate in the AI statement that a human who is fluent in both the original language and English has reviewed and assessed the translation quality, correcting any errors.

Like traditional sources, AI-generated works and information must be cited in an article when they are used as a source that is quoted, paraphrased, or otherwise incorporated. Any use of AI requires that the author(s) include the full output of the AI in supplemental material, other than its use for word selection and sentence-level editing.

Please refer to APA Publishing's AI policy as well, to which our journal also abides. Guidance is available at the link below on where within an article, information on AI use should be cited, although please keep in mind that we have further restrictions on AI use than APA has listed because of our journal’s specific focus on qualitative research.

Please note that AI cannot be an author or co-author on an article because it cannot be responsible for ensuring the accuracy and originality of the work. Human authors hold this responsibility and need to ensure that there is no plagiarism or errors in their text, images, citations, or references.

In the spirit of promoting transparency and openness, we encourage reaching out to the editor if author(s) have questions or require clarification. Articles that have been found to use AI in contravention of this statement may have their acceptances revoked. This position reflects our commitment to research ethics based in honoring the experiences and humanity of our participants and careful, in-depth analysis and interpretation of data.

We expect that this guidance will need to be updated periodically as information on AI use and practices in the field develop. Editors may consider individual submissions in light of changing practices and information.