Figure 1
PRISMA Chart.
Table 1
Distribution of AI-Use Disclosures by Journal.
| JOURNAL | NO. ARTICLES PUBLISHED IN STUDY PERIOD | NO. EMPIRICAL ARTICLES | NO. OF ARTICLE WITH DISCLOSURES | PERCENTAGE OF ARTICLES WITH DISCLOSURES IN THE JOURNAL | AI DISCLOSURE GUIDANCE AVAILABLE |
|---|---|---|---|---|---|
| Academic Medicine | 303 | 135 | 2 | 0.66% | Y |
| Advances in Medical Education and Practice | 94 | 80 | 0 | 0.00% | Y |
| Advances in Health Sciences Education | 81 | 63 | 1 | 1.23% | Y |
| Anatomical Sciences Education | 87 | 59 | 2 | 2.30% | N |
| BMC Medical Education | 856 | 796 | 24 | 2.80% | Y |
| Canadian Medical Education Journal | 72 | 27 | 0 | 0.00% | Y |
| Clinical Teacher | 137 | 97 | 1 | 0.73% | Y |
| Focus on Health Professional Education | 8 | 3 | 2 | 25.00% | Y |
| GMS Journal for Medical Education | 26 | 23 | 0 | 0.00% | N |
| International Journal of Medical Education | 14 | 13 | 0 | 0.00% | N |
| Journal of Continuing Education in the Health Professions | 41 | 30 | 1 | 2.44% | Y |
| Journal of Educational Evaluation for Health Professions | 16 | 16 | 0 | 0.00% | Y |
| Journal of Graduate Medical Education | 94 | 43 | 1 | 1.06% | Y |
| Journal of Medical Education and Curricular Development | 78 | 68 | 0 | 0.00% | Y |
| Journal of Surgical Education | 170 | 155 | 4 | 2.35% | Y |
| Medical Education | 170 | 77 | 0 | 2.25% | Y |
| Medical Education Online | 50 | 43 | 3 | 0.58% | Y |
| Medical Teacher | 304 | 153 | 7 | 2.30% | Y |
| Perspectives on Medical Education | 33 | 28 | 0 | 0.00% | Y |
| Simulation in Healthcare | 42 | 30 | 0 | 0.00% | Y |
| Teaching and Learning in Medicine | 62 | 44 | 1 | 1.61% | Y |
| Education for Health | 24 | 14 | 2 | 6.25% | Y |
[i] **African Journal of Health Professional Education and BMJ Stimulation & Technology Enhanced Learning did not have any empirical studies published to analyze.
Table 2
Regional Distribution (WHO Regions) of First Authors of Articles with AI-Use Disclosures (n = 51).
| REGION | NUMBER OF ARTICLES (%) |
|---|---|
| Western Pacific (e.g., Australia, China, South Korea) | 15 (29) |
| Americas (e.g., North, Central, South America) | 11 (22) |
| Europe (e.g., Germany, Netherlands) | 11 (22) |
| Eastern Mediterranean (e.g., Afghanistan, Egypt, Iran) | 9 (18) |
| South-East Asia (e.g., Bhutan, Bangladesh, India) | 4 (8) |
| Africa (e.g., South Africa) | 1 (2) |
| Total | 51 (100) |
Table 3
AI Tools Reported in Disclosure Statements (n = 51).
| AI TOOL | BRIEF DESCRIPTION OF TOOL | NUMBER OF ARTICLES (%)* |
|---|---|---|
| ChatGPT | Large Language Model | 19 (37%) |
| Otter | AI-Powered Transcription Software | 11 (22%) |
| Negative AI Disclosure | N/A | 5 (10%) |
| Unspecified | N/A | 4 (68%) |
| Editage | AI Editing for Researchers | 3 (6%) |
| Claude | AI Assistant | 2 (4%) |
| Curie | Clinical AI Assistant | 2 (4%) |
| ASReview | AI tool for Systematic Reviews | 1 (2%) |
| ClusterBot | AI systematic text categorization | 1 (2%) |
| Consensus.app | AI-powered Search Engine | 1 (2%) |
| Copilot | Generative AI assistant | 2 (2%) |
| DeepL | Comprehensive AI Language Platform | 1 (2%) |
| Grammarly | AI Writing Assistant | 1 (2%) |
| Gemini | Large Language Model | 1 (2%) |
| Hawk | Writing/reading AI Tool | 1 (2%) |
| Open Evidence | Evidence-based clinical decision making AI tool | 1 (2%) |
| POE AI Tools | AI Chatbot Aggregator | 1 (2%) |
| RapidMiner | Data Analytics AI Tool | 1 (2%) |
| Rev.com | Speech to Text Service | 1 (2%) |
[i] *Some articles reported multiple tools hence the total is greater than 51.
Table 4
Nature of AI Use in Disclosure Statements (n = 51).
| TYPE OF AI USE | DEFINITION | EXAMPLES | NUMBER OF ARTICLES (%) |
|---|---|---|---|
| Editing | Polishing | “Claude v. 3.5 Sonnet has been used for language editing. All content and ideas remain the original work of the authors, with AI assistance to improve linguistic clarity.” | 28 (55) |
| Transcription | Text generation from audio/video | “We used the Otter.ai transcribing software (Otter.ai Inc, California), with the RA reviewing each transcript for accuracy.” | 12 (24) |
| Analysis | Thematic and Statistical Analysis | “The authors acknowledge availing of a language quality checker and editing tool that used Curie’s AI software freely available from Springer Nature webpage as mentioned in the submission guidelines. The authors also acknowledge the use of Claude 3.5 Sonnet software for thematic analysis of the qualitative data.” | 7 (14) |
| Drafting | Text generation | “We acknowledge the assistance of the GPT-4 AI language model by OpenAI for support with statistical analysis and drafting of this manuscript” | 3 (6) |
| Literature Review: Article Screening | Includes article screening as a component of literature review | “We used an open-source artificial intelligence (AI) tool, ASReview (V.1.4) [21]. ASReview employs a machine learning algorithm that prioritizes articles based on their textual proximity to previously identified relevant articles (by the researchers). The tool consequently reduces the time and effort required during the initial screening phase but does not replace the initial screening of articles by researchers.” | 2 (4) |
[i] **Some articles cited multiple uses of AI hence the total being greater than 49.