Navigating the Landscape of Personalized Medicine: The Relevance of ChatGPT, BingChat, and Bard AI in Nephrology Literature Searches
Bibliographic record
Abstract
BACKGROUND AND OBJECTIVES: Literature reviews are foundational to understanding medical evidence. With AI tools like ChatGPT, Bing Chat and Bard AI emerging as potential aids in this domain, this study aimed to individually assess their citation accuracy within Nephrology, comparing their performance in providing precise. MATERIALS AND METHODS: We generated the prompt to solicit 20 references in Vancouver style in each 12 Nephrology topics, using ChatGPT, Bing Chat and Bard. We verified the existence and accuracy of the provided references using PubMed, Google Scholar, and Web of Science. We categorized the validity of the references from the AI chatbot into (1) incomplete, (2) fabricated, (3) inaccurate, and (4) accurate. RESULTS: A total of 199 (83%), 158 (66%) and 112 (47%) unique references were provided from ChatGPT, Bing Chat and Bard, respectively. ChatGPT provided 76 (38%) accurate, 82 (41%) inaccurate, 32 (16%) fabricated and 9 (5%) incomplete references. Bing Chat provided 47 (30%) accurate, 77 (49%) inaccurate, 21 (13%) fabricated and 13 (8%) incomplete references. In contrast, Bard provided 3 (3%) accurate, 26 (23%) inaccurate, 71 (63%) fabricated and 12 (11%) incomplete references. The most common error type across platforms was incorrect DOIs. CONCLUSIONS: In the field of medicine, the necessity for faultless adherence to research integrity is highlighted, asserting that even small errors cannot be tolerated. The outcomes of this investigation draw attention to inconsistent citation accuracy across the different AI tools evaluated. Despite some promising results, the discrepancies identified call for a cautious and rigorous vetting of AI-sourced references in medicine. Such chatbots, before becoming standard tools, need substantial refinements to assure unwavering precision in their outputs.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.118 | 0.590 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.035 | 0.026 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.009 | 0.013 |
| Open science | 0.002 | 0.008 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.006 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".