Methods for using Bing's AI-powered search engine for data extraction for a systematic review

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Systematic Review Methodology

Background:

• Data extraction in systematic reviews is labor-intensive.
• Natural Language Processing (NLP) and Artificial Intelligence (AI) offer automation potential.
• Enhancing efficiency and reliability in systematic reviews is crucial.

Purpose of the Study:

• To propose and demonstrate a method using Bing AI as a secondary reviewer for data extraction.
• To evaluate the potential of AI in verifying and enhancing data extracted by human reviewers.
• To provide a cost-effective solution for resource-limited systematic reviews.

Main Methods:

• A worked example detailing the use of Bing AI Chat to extract study characteristics from PDF documents.
• Instructing AI to populate data into a table for comparison with human-extracted data.
• Utilizing Microsoft Edge as a platform for AI-assisted verification.

Main Results:

• Bing AI can be instructed to extract specific data items from documents.
• The AI-assisted method offers an additional layer of verification for data extraction.
• This approach may be beneficial for reviewers with limited resources or experience.

Conclusions:

• Bing AI shows promise as a supplementary tool for data extraction in systematic reviews.
• It can enhance verification processes, especially when resources are scarce.
• Further research is needed to validate AI's accuracy and efficiency against established double-extraction methods.