RAISE Framework
Responsible Automated in
Evidence Synthesis
BioClaritas is committed to the responsible use of artificial intelligence in systematic reviews and evidence synthesis. We follow the RAISE framework to ensure all Automated-assisted outputs are transparent, validated, and ethically produced.
Core Principle
Evidence synthesis authors are ultimately responsible for the integrity of their work, including any outputs generated or assisted by Automated tools. Automated should augment — never replace — expert human judgement in the evidence synthesis process.
1Recommendations for Evidence Synthesis Authors
1.1 Evaluate Whether Automated Tools Are Appropriate
- •Before using Automated-based tools, assess the appropriateness, accuracy, and limitations for each specific evidence synthesis task.
- •Consider whether the Automated tool has been validated for the specific task context (e.g., screening, data extraction, risk-of-bias assessment, summarisation).
- •Understand the tool's training data, known biases, and error rates to make informed decisions about when and how to use it.
1.2 Maintain Human Oversight and Validation
- •Always verify Automated-generated outputs through expert human review. Automated outputs should be treated as a starting point, not a final product.
- •Implement quality assurance workflows that include independent human validation of Automated-assisted steps.
- •Document any discrepancies between Automated outputs and human reviewer judgements, and describe how conflicts were resolved.
1.3 Report Automated Use Transparently
- •Clearly report all Automated tools used in the methods section of your manuscript, including the tool name, version, provider, and underlying model.
- •Describe exactly which steps of the evidence synthesis process used Automated assistance (e.g., title/abstract screening, full-text screening, data extraction, risk-of-bias assessment, narrative synthesis, illustration generation).
- •Report the prompts, parameters, and settings used so that reviewers and readers can assess and reproduce the work.
- •Automated tools should not be listed as authors. Acknowledge their use in the methods or an acknowledgements section instead.
1.4 Ensure Ethical, Legal, and Regulatory Compliance
- •Ensure that the use of Automated tools complies with institutional ethics policies, data protection regulations (e.g., GDPR, HIPAA), and applicable Automated regulations.
- •Do not upload sensitive, personally identifiable, or proprietary data to Automated tools unless explicitly permitted and appropriately safeguarded.
- •Be aware of intellectual property considerations — understand the licensing terms of the Automated tools and how outputs may be used or shared.
1.5 Contribute to the Evidence Synthesis Automation Ecosystem
- •Where possible, share validation studies, benchmarks, and best practices with the broader evidence synthesis community.
- •Provide feedback to Automated tool developers about shortcomings, errors, or areas for improvement.
- •Advocate for open science principles — share protocols, data, and code when feasible to enable reproducibility.
2Recommendations for Funders
- •Fund research into the validation and evaluation of Automated tools for evidence synthesis tasks.
- •Support the development of open-source, transparent Automated tools that are accessible to researchers globally.
- •Require funded projects that use Automated in evidence synthesis to adhere to RAISE guidelines and report Automated use transparently.
- •Invest in training programmes that equip evidence synthesists with the skills to use Automated tools responsibly.
3Recommendations for Publishers
- •Develop and enforce clear editorial policies on the use of Automated in evidence synthesis manuscripts.
- •Require authors to declare all Automated tool usage and provide sufficient detail for peer reviewers to evaluate the methods.
- •Train peer reviewers to evaluate Automated-assisted evidence syntheses critically, including assessing validation and transparency of Automated use.
- •Consider RAISE-aligned reporting checklists as part of the submission and review process.
4Recommendations for Users of Evidence Syntheses
- •When reading Automated-assisted evidence syntheses, critically appraise how Automated tools were used and whether appropriate validation was performed.
- •Look for transparent reporting of Automated tool names, versions, specific tasks, and human validation steps.
- •Be cautious about evidence syntheses that rely heavily on Automated without adequate human oversight or fail to report Automated use.
- •Consider the potential for Automated-introduced biases when interpreting findings from Automated-assisted reviews.
5Recommendations for Trainers & Educators
- •Integrate Automated literacy and responsible Automated use into evidence synthesis training curricula.
- •Teach trainees to critically evaluate Automated tools, understand their limitations, and implement appropriate validation workflows.
- •Emphasise that Automated tools are aids to the evidence synthesis process, not substitutes for methodological expertise and critical thinking.
- •Provide practical exercises on validating Automated outputs, identifying errors, and reporting Automated use in publications.
How BioClaritas Implements RAISE
Human-in-the-Loop
All Automated-generated outputs require human review and validation before use.
Transparent Processing
We clearly document which Automated models power each feature so you can report them accurately.
Data Privacy
Your research data is never used to train Automated models. We process data only for your requested tasks.
Editable Outputs
Every Automated-generated extraction, summary, and illustration can be reviewed, edited, and corrected by the user.
Your Responsibilities as a User
By using BioClaritas, you acknowledge and agree that:
- You are ultimately responsible for validating all Automated-generated outputs before incorporating them into your research.
- You will transparently report your use of BioClaritas and any Automated tools in your publications, following RAISE and applicable reporting guidelines (e.g., PRISMA).
- You will not list Automated tools as co-authors but will acknowledge their use appropriately.
- You will ensure your use of Automated tools complies with your institutional policies, journal requirements, and applicable data protection regulations.
- You will not upload sensitive or personally identifiable data unless explicitly permitted and safeguarded.
- You understand that Automated can introduce errors and biases, and you will implement quality assurance measures accordingly.
Citing RAISE
The RAISE framework is developed by the evidence synthesis community to promote responsible Automated use. When referencing these guidelines in your publications, please cite the original RAISE publication and mention your use of BioClaritas in the methods or acknowledgements section.
Suggested reporting text: “Automated-assisted evidence synthesis was conducted using BioClaritas (bioclaritas.com), following the Responsible Automated in Evidence Synthesis (RAISE) guidelines. All Automated-generated outputs were validated by [number] independent human reviewers.”