AI evidence extraction

AI TOOLS FOR LITERATURE REVIEW AND KNOWLEDGE MAPPING: EMPOWERING RESEARCH EXCELLENCE IN ACADEMIC WRITING AND PUBLISHING

Ramakrishna Rao Pulletikurty · Journal of Emerging Technologies and Innovative Research · 2026

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

9/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.56975/jetir.v13i2.575657

Methodology & findings

Study design

Qualitative literature review with thematic content analysis.

Primary method

No quantitative statistical methods employed. The study used qualitative thematic content analysis and conceptual synthesis of published evidence and expert opinions.

Main result

The study found that AI tools can "save researchers considerable time and help ensure that critical studies are not overlooked" through capabilities including semantic search, automated summarization, and knowledge mapping. However, the authors emphasize that "the best outcomes arise when AI is used as an intelligent aid, not a replacement for critical thinking," and note that Meliante et al. (2025) found AI tools "did not retrieve the full set of relevant studies compared to manual methods."

Reports effect sizes.

Research paradigm

Interpretivist/qualitative synthesis

Author conclusions

The authors conclude that "AI holds great promise for advancing research quality and impact. When integrated responsibly-with proper training, policy, and human collaboration-AI can significantly empower academic writing and publishing." They emphasize that "the best outcomes arise when AI is used as an intelligent aid, not a replacement for critical thinking" and that "ethical awareness and oversight are essential: researchers should ensure that AI-assisted findings are validated and that all contributors (human and machine) are appropriately credited."

Risk of bias

Algorithmic bias - AI models may encode biases from training data (over-representation of certain fields, languages, or demographics); Language bias - search limited to English-language publications; Field bias - sources primarily from library and information science, computer science, education, and publishing domains; Transparency/black box problem - many AI methods lack interpretability; Publication bias - relies on published evidence and reports; Algorithmic bias encoding over-representation of certain fields, languages, or demographics; Data privacy concerns when researchers upload sensitive or proprietary documents to AI platforms; Lack of transparency in deep learning models ('black boxes'); Selection bias in search results based on training data limitations; Algorithmic bias in AI systems reflecting training data limitations and over-representation of certain fields, languages, or demographics; Selection bias in the literature review process (English-language publications only, primarily from specific domains); Black box opacity of AI methods making it difficult to understand recommendation and summarization processes; Publication bias toward positive portrayals of AI tools in academic and industry reports

Limitations

The authors identify several limitations through their analysis of AI tools: "AI tools tested for literature reviews did not retrieve the full set of relevant studies compared to manual methods," and "the active participation of the researcher ..
is still crucial to maintain control over the quality, accuracy, and objectivity" of the review
Additionally, "current AI systems sometimes produce incorrect or nonsensical outputs (often called 'hallucinations')" and "studies have shown accuracy around only 50-60% for some tools." The authors also note that "AI models learn from existing data, which can encode biases (e.g
over-representation of certain fields, languages, or demographics)."

Open questions raised

Need for empirical evaluations of specific AI tools in literature review workflows
Lack of standardized guidelines for responsible AI use in academia
Insufficient training programs on AI literacy for researchers and students
Gaps in understanding full scope of algorithmic bias in AI discovery systems
Limited research on skill erosion effects from over-reliance on AI tools
The paper identifies the need for: (1) updated academic guidelines addressing AI ethics and authorship; (2) strengthened human-AI partnerships with human-in-the-loop workflows; (3) increased AI literacy training in universities and research institutions; (4) institutional policies on AI use including citation norms and intellectual property standards; (5) enhanced role of libraries and information professionals in curating AI tools and teaching best practices.

Extracted from: pdfAgreement 66%

Explore related topics

Systematic Review Automation Literature Discovery Literature Review