Comportement des agents conversationnels quant aux rôles et retombées du pharmacien : une étude exploratoire
Bibliographic record
Abstract
Objectif : Décrire le profil général du comportement de deux agents conversationnels et leurs différentes versions existantes et qualifier l’adéquation des réponses proposées à des questions sur les rôles et retombées du pharmacien. Qualifier la justesse des références bibliographiques proposées par ces agents conversationnels. Méthode : Étude descriptive et qualitative de type transversal exploratoire. Deux agents conversationnels ont été sélectionnés, soit ChatGPT (version 3.5 gratuite, version 4 payante) et Bing (trois versions, soit Bing équilibré, précis et créatif). Quarante-six questions sur les rôles et les retombées de l’activité pharmaceutique ont été formulées. Un panel de trois experts a été formé afin de juger de la justesse des réponses. De plus, la qualité des références a été décrite et évaluée. Seules des statistiques descriptives ont été effectuées. Résultats : Les deux agents utilisés proposent une proportion très élevée de réponses jugées adéquates par le panel d’experts, soit 95 % (ChatGPT-3.5 et ChatGPT-4), 91 % (Bing équilibré), 98 % (Bing créatif) et 91 % (Bing précis). La proportion de références adéquates varie d’un agent à l’autre, soit 75 % (3/4, ChaptGPT-3.5), 80 % (4/5, ChatGPT-4), 89 % (48/54, Bing équilibré), 91 % (133/146, Bing créatif) et 91 % (91/100, Bing précis). Conclusion : Cette étude montre que deux agents conversationnels (Chat GPT-3.5, ChatGPT-4, Bing équilibré, précis, créatif) proposent des réponses adéquates à une série de questions entourant les rôles et retombées des pharmaciens. Toutefois, ChatGPT propose un nombre plus limité de références que Bing et parfois des références inadéquates ou inventées. Il apparaît donc nécessaire de continuer d’évaluer l’utilité et la justesse des agents conversationnels en pharmacie. Summary Objective: To describe the general behavioural profile of two chatbots and their different existing versions and to assess the adequacy of proposed answers to questions about the role and impact of pharmacists. To assess the accuracy of the bibliographic references provided by these chatbots. Method: Descriptive and qualitative exploratory cross-sectional study. Two chatbots were selected: ChatGPT (free version 3.5, paid version 4) and Bing (three versions, namely Bing balanced, precise, and creative). Forty-six questions about the role and impact of pharmacist-led activities were developed. A panel of three experts was formed to assess the accuracy of chatbot answers. In addition, the quality of the references provided by the chatbots was described and evaluated. Descriptive statistics were conducted. Results: The two chatbots provided a very high proportion of responses deemed adequate by the panel of experts, i.e. 95% (ChatGPT-3.5 and ChatGPT-4), 91% (Balanced Bing), 98% (Creative Bing) and 91% (Precise Bing). The proportion of adequate references varied from one chatbot to another, i.e. 75% (3/4, ChaptGPT-3.5), 80% (4/5, ChatGPT-4), 89% (48/54, Balanced Bing), 91% (133/146, Creative Bing) and 91% (91/100, Precise Bing). Conclusion: This study shows that two chatbots (Chat GPT-3.5, ChatGPT-4, Bing balanced, precise, creative) provide adequate answers to a series of questions surrounding the role and benefits associated with pharmacist-led activities. However, ChatGPT offers a more limited number of references than Bing and sometimes offers inadequate or invented references. It therefore seems necessary to continue to evaluate the usefulness and accuracy of chatbot use in pharmacy.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.008 | 0.000 |
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 teacher head, 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".