AI evidence extraction

Empowering biological knowledgebases: advances in human-in-the-loop AI-driven literature curation

Valerie Wood, Matt Jeffryes, Andrew F Green, Matthias Blum, Sandra Orchard, Simona Panni et al. · Bioinformatics Advances · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

0/4

Quality (LMQS)

Evidence

Citations

13.36

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1093/bioadv/vbag028

Methodology & findings

Study design

Narrative review of literature and published use cases at Global Core BioData Resources and ELIXIR Core Data Resources, synthesizing evidence on AI applications in biological biocuration

Main result

The paper identifies that "artificial intelligence, particularly large language models and agentic systems, can augment literature-curation workflows" through applications including "literature recommendation, entity recognition, data extraction, summarization, ontology development, and quality control." The authors emphasize that a "human-in-the-loop framework where generative artificial intelligence approaches accelerate routine tasks while curators provide critical evaluation and domain expertise" is essential for maintaining biological rigour.

Research paradigm

Interpretivist/Qualitative - Critical examination of AI applications in biocuration with emphasis on human-in-the-loop frameworks

Author conclusions

The authors conclude that "These synergistic partnerships will be critical to ensure biological rigour, accelerating knowledge integration while maintaining the quality essential for trusted biological resources." They propose that the community should focus on "the creation of shared benchmark datasets, harmonized evaluation frameworks, and best-practice guidelines for transparent human-in-the-loop AI deployment in biocuration."

Risk of bias

Selection bias in reviewed use cases (only published, successful implementations may be highlighted); Funding bias potential (review emphasizes AI-augmented approaches which may align with funding priorities); Coverage bias (focus on Global Core BioData Resources and ELIXIR may not represent broader biocuration landscape)

Limitations

The paper identifies key challenges including "the scarcity of training data, difficulty in extracting complex relationships, and concerns about error propagation." The authors note that biological knowledgebases are "maintained by a small number of professional biocurators worldwide and face combined chronic underfunding and the exponential growth of the literature."

Open questions raised

The authors identify the need for: (1) shared benchmark datasets for biocuration tasks; (2) harmonized evaluation frameworks for assessing AI-assisted curation; (3) best-practice guidelines for transparent human-in-the-loop AI deployment in biocuration; (4) solutions to address scarcity of training data; (5) improved methods for extracting complex relationships in biological literature; (6) strategies to mitigate error propagation in AI systems used for knowledge curation.
The paper identifies the need for: (1) shared benchmark datasets for biocuration AI systems, (2) harmonized evaluation frameworks across biocuration tools, (3) best-practice guidelines for transparent human-in-the-loop AI deployment, and (4) solutions to address scarcity of training data and error propagation concerns
The authors identify the need for: (1) shared benchmark datasets for AI evaluation in biocuration; (2) harmonized evaluation frameworks across different biocuration tasks; (3) best-practice guidelines for transparent human-in-the-loop AI deployment; (4) solutions to address scarcity of training data; (5) improved methods for extracting complex biological relationships; and (6) mechanisms to prevent error propagation in AI-assisted curation.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 71%

Explore related topics

Systematic Review Automation Literature Discovery Literature Review