academic writing, apa writing standards, and artificial intelligence

Geese writing papers, scientists and regulators observing

When writing, you should follow an academic writing system—I recommend following the APA 7th edition publication manual—unfortunately, the rules about the use of artificial intelligence are contradictory. Wherever you stand on the debate around artificial intelligence—its impact on the environment and local communities, and its use in academia and professional writing—we should all agree that it is dangerous for the future of professional writing when institutional rules incentivize deception, undisclosed automation, uncritical sourcing, and the intellectual dishonesty of passing off machine output as human thought.

Not all uses of artificial intelligence are harmful; when used thoughtfully, AI can meaningfully accelerate the mechanical sides of writing and editing. The danger arises when writers outsource the core generation of ideas to a model rather than synthesizing research through their own critical judgment, eroding both the integrity and authority of their work. If there is one critical takeaway, it is that any use of artificial intelligence—including specific tools like ChatGPT, NotebookLM, or Quillbot—must always be transparently disclosed alongside an explanation of how it was applied.

At the same time, maintaining this standard requires a supportive ecosystem. Unrealistic citation requirements and a lack of practical tracking tools currently exacerbate the problem for writers trying to do the right thing. Professional writing standards must adapt, giving writers accessible, practical tools so they can meet transparency requirements with clarity rather than friction.

This blog post reviews the guidance provided by the Publication Manual of the American Psychological Association (7th Edition), along with additional recommendations from the American Psychological Association and other institutional sources.

contradictory and inconsistent guidance from the american psychological association

Let’s take a look at the American Psychological Association’s guidelines on the use of AI in professional writing, which are best described as contradictory and inconsistent and dispersed across multiple unrelated webpages with occasionally context-dependent rules.

The American Psychological Association (2026) states in its APA Style wiki guidelines that artificial intelligence should be cited whenever it is used. The guidelines include examples for citing specific AI tools and individual chats. Here are my reflections on this guidance.
- Citing the tool itself instead of the specific chat appears misleading. In professional writing, it is essential to provide empirical sources rather than relying on vague assertions like, “Trust me.” Citing ChatGPT as a source is akin to referencing Google instead of the specific webpage from which the information originated; it would be unacceptable to list Google as a source in an academic paper, and I believe the same standard should apply to ChatGPT.
- I still need to ask, “Where did the information come from?” even if we cite the particular AI chat and treat it as a tertiary aggregator, which it might not be eligible for. You still have to cite books even if you know them all by heart. The fact that an AI wrote the response does not make it authoritative or alter the rules.
In a recent article published on the APA Style blog, McAdoo et al. (2025) explain that AI cannot be an author or have authorship. When publishing a professional paper, you cannot list AI as an author. The APA does not recognize AI as an author; using AI to generate information and then claiming you wrote it is unethical. Citing information from AI may conflict with the APA’s guidelines on authorship, even if you attempt to reference a specific chat. This raises an important question: Why write citation guidelines for AI when the APA’s own rules prohibit citing AI as an author? It’s impossible to follow both standards at the same time.
The APA provides even stricter guidance for journal articles. The American Psychological Association (2025) explains that generative AI cannot be an author; the APA’s ethical standards and code of conduct must be followed. Generally, AI chats must be preserved for inquiry, and the tool and prompts must be disclosed in the paper. Do I have to mention in my research paper’s introduction that I used the AI writing and editing tool Quillbot?
According to McAdoo et al. (2025), Microsoft Copilot is not required to be cited within Microsoft Word. I hope this only affects the editing and phrase features, but the chat sidebar might also be affected. Can I avoid disclosure requirements if I only use AI within my word processor rather than the website? I hope not!

It is impossible to comply with all of the APA’s guidance around AI if you use AI at all! Given the inability to follow APA guidance while using AI, will researchers avoid using it? I suspect that AI use is not disclosed or cited in practice.

what does academic research say?

According to BaHammam (2025), while the use of AI has increased in scientific writing, the citations and disclosures are not. The research paper breaks this issue down into four categories: The stigma of using AI in academia, the productivity these tools provide, systemic issues (such as the APA guidance inconsistency), and cultural and institutional factors. The research discusses a survey of researchers about their use of AI and their disclosures. It found that 28% of researchers admit to using AI, 8% generated their first drafts, and the majority of people who admitted to using AI did not disclose their use of AI at all. Most importantly the research article proposed a framework for how AI use should be disclosed.

what is the impact of not citing or disclosing ai use?

When researchers reading a paper do not know that AI was used in the drafting process of a research paper, they cannot determine how reliable the research is. Citing ChatGPT as a source is like citing Google and not the actual web page the data came from.
While it is safe to choose not to cite or disclose the use of AI because it is difficult to prove and AI-generated research is published, people who attempt to abide by the current regulations and disclose the use of AI are frequently stigmatized or punished.
The public’s confidence in the validity of research papers and the peer-review procedure is impacted.
When false information is published and passes peer review, it is disseminated by others who begin citing the research paper. It is unlikely that the paper will be removed from publication.

next steps

Institutions must assess the risks associated with artificial intelligence and offer clear guidance on the ethical application of these tools.
Institutional guidance should be consistent across all platforms, including websites and databases. Inconsistencies across multiple webpages can confuse good faith actors and hinder compliance.
In order to make organizing the many chats involved in a project easier to locate in the event of an inquiry, AI platforms such as Claude, NotebookLM, and Grammarly should offer immutable chats, search tools, and organizational features.
In order to create a framework that is uniform across all writing standards regarding the use of AI in professional writing, writing standards organizations (like APA, MLA, and Chicago) must work together. Reducing compliance nightmares and spending less time in doubt about small differences between standards.

references

American Psychological Association. (2026, April). AI references. https://apastyle.apa.org/style-grammar-guidelines/references/examples/ai-references

American Psychological Association. (2025, August). APA Journals policy on generative AI: Additional guidance. https://www.apa.org/pubs/journals/resources/publishing-tips/policy-generative-ai

BaHammam A. S. (2025). The Transparency Paradox: Why Researchers Avoid Disclosing AI Assistance in Scientific Writing. Nature and science of sleep, 17, 2569–2574. https://doi.org/10.2147/NSS.S568375

McAdoo, T., Denneny, S., & Lee, C. (2025, September 9). Citing generative AI in APA Style: Part 3 – Is AI “allowed” in APA Style? APA Style Blog. https://apastyle.apa.org/blog/cite-generative-ai-allowed

McAdoo, T., Denneny, S., & Lee, C. (2025, September 9). Citing generative AI in APA Style: Part 2 – AI as a search engine and AI integrated into common software. APA Style. https://apastyle.apa.org/blog/cite-generative-ai-search-software