CONTENT AND QUALITY OF CHATGPT AS A SOURCE OF DEMENTIA INFORMATION: COMPARATIVE ANALYSES
Bibliographic record
Abstract
Abstract There is significant demand for easily accessible, reliable online health information from older adults and their families (Soong et al., 2020). Online information about Alzheimer’s Disease and dementia ranges in quality from misleading and predatory to supportive and evidence-based (Robillard & Feng, 2016). ChatGPT was released to the public in November 2022 and is a well-known example of a new class of online tools, Large Language Models, which are generative artificial intelligence systems that can produce fluent, coherent answers to users’ questions based on extensive training data. In this work, we use a validated evaluation tool (Robillard et al., 2018) to evaluate the quality and content of information about dementia from ChatGPT-3.5 at two time points (2023, 2024) and from two versions of the model (free GPT-3.5, paid GPT-4). Prompts were developed using questions extracted from the FAQ pages of national dementia organization websites in Canada, USA, and Mexico. We found that all versions of ChatGPT were unbiased, remind the user to speak to a physician, and took an appropriately balanced tone. GPT-4, unlike GPT-3.5, links to specific, identifiable, dated sources for its claims including scientific research. All forms of ChatGPT had a high Flesch-Kincaid Grade Level, indicating moderate readability. Results indicate that ChatGPT, particularly GPT-4, can be a relatively high-quality source of dementia information for non-expert users. However, it is limited in its ability to link users to local resources. Findings can inform older adults, care partners, and healthcare professionals in their decision-making around use of this new tool.
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 | 0.021 | 0.181 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.003 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.004 | 0.004 |
| Open science | 0.001 | 0.003 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 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".