AI Policy

Role of Artificial intelligence

Artificial intelligence (AI) technologies are used in the CAM Cancer process as supportive tools to enhance efficiency and consistency in selected methodological steps. AI does not replace human judgement: all AI-assisted outputs are reviewed and validated by CAM Cancer’s experienced authors, reviewers and editors.

AI-Assisted Tasks

Study screening and management 

Rayyan, an AI-assisted screening tool, is used to manage references and support title and abstract screening. CAM Cancer uses the platform as it facilitates collaboration and enables efficient application of inclusion and exclusion criteria. Its features such as e.g. ranking and inclusion/exclusion suggestions assist authors but do not determine final decisions.

Study selection and eligibility

AI is used to support the initial assessment of methodological quality of systematic reviews (SRs) using the AMSTAR2 tool.

• An AI tool (AI assistant M365 Copilot, Microsoft) is used as a supportive tool to generate preliminary AMSTAR 2 assessments based on predefined criteria (Shea 2017). The tool does not make final quality judgements.
• One author evaluates and revises all AI-generated assessments to establish whether a SR is of sufficient methodological quality to be included.
• A second author checks the final AMSTAR2 ratings to ensure methodological accuracy and consistency. Responsibility for all AMSTAR 2 ratings rests with the authors.

Data Extraction

Data extraction is conducted by authors with assistance from an AI tool (M365 Copilot, Microsoft). The tool is used to support identification and structuring of pre-specified data elements from included studies. All extracted data are independently verified by a human reviewer against the original publications and spot-checked by the second author.

• Study characteristics (e.g. population, cancer type, interventions, controls)
• Outcomes assessed
• Main results for risk of bias assessment
• Main results for effects with effect sizes
• Authors’ verbatim conclusions

Quality Assurance and Transparency

• All final decisions, including AI-assisted study eligibility, quality and extracted data, reflect author judgement.
• AI-assisted outputs are treated as draft inputs only and no automated decisions are made.
• Discrepancies or uncertainties identified in AI-assisted outputs are resolved through manual review of the original publications.

Limitations

The use of AI may be limited by the quality, reporting clarity, and accessibility of the included publications. CAM Cancer observed that tables and graphs are prone to errors when AI tools are used for data extraction; consequently, these tasks are always performed by human authors. AI tools may misinterpret complex methodological details; therefore, human oversight is used at all stages where AI is applied.

Ethical and Methodological Considerations

The use of AI in CAM Cancer summaries aims to enhance reproducibility and efficiency while maintaining methodological rigour. AI is used in line with principles of transparency, accountability, and author responsibility.

Conflict of interest

The CAM Cancer editorial team has no financial or non-financial interests related to the AI tools it uses. Individual authors are requested to disclose any potential conflicts of interest (COI) in accordance with our COI policy on identifying, preventing, or managing COI.