What Does It Take to Research with AI? A Rapid Review of Competencies to Train LLM-Literate Researchers

2026-07-20 · Source: cs.SE updates on arXiv.org · Field: Science & Research — Research Methodology & Innovation, Engineering & Applied Sciences, Social Sciences & Behavioral Studies · Depth: Intermediate, extended

Summary

A rapid review of 194 articles published between 2022 and 2025, sourced from Elicit and Google Scholar, identified eight critical competencies for researchers and graduate students using Large Language Models (LLMs) in scientific research. After dual screening 40 selected articles (Gwet AC1: 0.76–0.83), the most prevalent competency was "domain expertise and oversight of AI outputs" (Σn=123), emphasizing subject-matter mastery, systematic skepticism, source verification, and researcher accountability. Other key competencies include "metacognition and decision making about AI use" (Σn=55), "ethics and academic integrity" (Σn=53), "prompt engineering for research" (Σn=38), and "reproducibility of AI use" (Σn=29). "AI literacy and technical knowledge" (Σn=16) was noted as a risk when absent, with domain expertise being crucial for critical evaluation. The study concludes that preparing researchers for LLM use requires an integrated set of epistemic, ethical, and methodological skills, focusing on human accountability, with direct implications for graduate programs and AI literacy initiatives.

Key takeaway

For AI students and research scientists integrating LLMs into your work, recognize that effective use extends beyond technical proficiency. You must prioritize developing strong domain expertise, critical oversight of AI outputs, and robust ethical decision-making. Document your AI usage transparently, including prompts and model versions, to ensure reproducibility and maintain accountability for all research conclusions. Your judgment remains paramount in validating AI-generated content.

Key insights

LLM-assisted research requires integrated human competencies, prioritizing accountability, critical judgment, and ethical oversight over mere technical skills.

Principles

• Domain expertise is foundational for evaluating AI outputs.
• Researchers must maintain human oversight and accountability for AI-generated content.
• Ethical considerations are integral throughout the AI-assisted research process.

Method

A rapid review analyzed 194 articles (2022–2025) from Elicit and Google Scholar. 40 articles were selected via dual screening (Gwet AC1: 0.76–0.83) for thematic analysis, identifying and consolidating eight key competencies for LLM-assisted research.

In practice

• Verify AI-generated claims against original sources.
• Document prompts, model versions, and AI-assisted tasks.
• Design research workflows accounting for AI strengths and limitations.

Topics

• Large Language Models
• Research Competencies
• AI Literacy
• Prompt Engineering
• Academic Integrity
• Research Reproducibility

Best for: AI Scientist, Research Scientist, AI Student, AI Ethicist

Related on AIssential

• Attributing and situating knowledge cannot be left to language models — LLMs cannot reliably handle scholarly attribution, risking miscitation and undermining research standards.
• We Need to Talk About AI — AI is rapidly transforming scientific research, raising profound questions about human intellectual roles and the future of discovery.
• Mar 23, 2026ScienceVibe physics: The AI grad student — AI can significantly accelerate expert-guided theoretical physics research, but requires rigorous human oversight.
• Chaptered Recording: The Past, Present, and Futures of Pragmatic and Responsible Adoption of AI in… — AI, particularly LLMs, offers comprehensive automation for systematic reviews, requiring pragmatic adoption and careful error management across all stages.
• AI and the PhD student: friend or foe? — AI tools offer efficiency gains for researchers but necessitate rigorous human verification to prevent skill erosion and factual errors.
• (Human) Attention Is (Still) All You Need: Human oversight makes AI-assisted social science reliable — Human oversight and structured decision architecture are crucial for reliable AI-assisted social science research, significantly reducing LLM-driven failures.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.SE updates on arXiv.org.