AI Disclaimer
Robyn Dawes Institute for the Improvement of Science provides AI Systems, tools, and products designed to extract and organize structured information from scientific publications at scale and to apply statistical methods commonly used in research analysis. These tools are hosted through https://evidence.guide and are provided for research, analysis, and informational support.
AI Limitations and User Responsibility
The Site, Services, and AI Systems do not replace independent scientific judgment, specialized expertise, or the nuanced contextual understanding of qualified professionals. Users remain solely responsible for analyzing and evaluating any AI-generated content to ensure it aligns with their needs, institutional requirements, and applicable standards.
Outputs are generated through automated extraction and statistical modeling processes and may be incomplete, probabilistic, or context-dependent. Results are based on statistical patterns and large-scale data processing and may not always be accurate or suitable for every context. Users should independently verify information and evaluate suitability before relying on any output.
The Site, Services, and AI Systems are provided on an "as is" and "as available" basis, without representations or warranties of any kind, whether express or implied, including regarding accuracy, completeness, reliability, or fitness for a particular research purpose.
Outputs generated by the AI tools are produced through automated processes and may not qualify for copyright protection, depending on the level of human authorship involved. Robyn Dawes Institute makes no representation or warranty regarding the ownership, originality, or protectability of any AI-generated content.
Dataset Transparency
Users should be aware that datasets used in developing, testing, or operating the Services and AI Systems may include a variety of materials and sources:
- Sources/owners: The primary data sources are user-uploaded scientific academic articles. Examples of publishers may include Elsevier, SAGE, Wiley, Nature Portfolio/Springer, and Oxford University Press.
- Volume/data points: The Services may extract and process structured information from academic articles, including numerous extracted signals per article.
- Types of data: Academic article PDFs and supplementary materials; extracted structured fields derived from those materials.
- Intellectual property: Academic articles are generally copyrighted. We do not expect the datasets to include trademarked or patented materials.
- Personal information: We do not expect the raw data to generally contain personal information or aggregate consumer information.
- Processing: PDFs may undergo preprocessing and transformation to convert materials into structured data for analysis.
- Time period: No specific time period for collection or use is currently defined.
- Synthetic data: No synthetic or AI-generated data is used at this time.
Datasets may include peer-reviewed academic journals, scientific and educational publications, research repositories, and other scholarly materials originating from publicly available materials, licensed datasets, publisher-provided content, or other rights-compliant frameworks where applicable.
Nothing in this section should be interpreted as a representation or warranty regarding completeness, ownership, licensing status, or accuracy of any dataset.
Model Development and Performance Limitations
Certain product features rely on automated extraction of structured information from scientific publications and the application of established statistical procedures used in research analysis.
Outputs may contain extraction errors, statistical limitations, or other inaccuracies, and performance may vary depending on inputs, context, and ongoing system development. Any references to validation methods, performance metrics, or accuracy rates are illustrative only and do not constitute representations, warranties, or guarantees of performance.
Development practices may reference recognized technical or risk-management frameworks (such as those published by NIST) for informational purposes only, and do not constitute certification, endorsement, or formal compliance unless expressly stated.
Questions
If you have questions about the nature of Robyn Dawes Institute services or any of our terms or policies, please contact us at [email protected].
Last Modified: April 2, 2026.