Attributing and situating knowledge cannot be left to language models

2026-02-06 · Source: Nature Machine Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Advanced, quick

Summary

The article "Attributing and situating knowledge cannot be left to language models" by Radu and Rocher, published in Nat Mach Intell on February 6, 2026, argues that meticulous citation is crucial for scholarly integrity, situating claims, and enabling scrutiny. Careless citation, such as overlooking original sources, miscredits ideas and reveals limited scholarship, disproportionately harming underrepresented voices whose work is often more innovative but less cited. The authors highlight that research led by women and underrepresented groups tends to be less cited. While large language models (LLMs) are increasingly used by students and researchers to draft and edit articles, potentially accelerating research, their reliance risks undermining established standards for research excellence.

Key takeaway

For AI Scientists and researchers integrating LLMs into their writing workflow, you must implement rigorous manual verification of all citations. Relying solely on LLM-generated references risks perpetuating citation biases, miscrediting original work, and undermining the credibility of your scholarship. Prioritize independent source verification to maintain research excellence and ethical attribution standards.

Key insights

LLMs cannot reliably handle scholarly attribution, risking miscitation and undermining research standards.

Principles

• Meticulous citation ensures scholarly integrity.
• Misattribution harms underrepresented researchers.

In practice

• Review LLM-generated citations carefully.
• Verify original sources for all references.

Topics

• Large Language Models
• Citation Practices
• Research Ethics
• Academic Scholarship
• Research Bias

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

Related on AIssential

• Impact of large language models on peer review opinions from a fine-grained perspective: Evidence from top conference proceedings in AI — LLMs are making peer reviews longer and more fluent but reducing focus on deep evaluative dimensions.
• PeerPrism: Peer Evaluation Expertise vs Review-writing AI — Current LLM detectors conflate text generation with intellectual contribution in hybrid human-AI peer review.
• Trust, but Don't Verify: Epistemic Blind Spots in LLM Source Evaluation — LLMs can detect fabricated statistics but fail to apply this capability during multi-source synthesis, prioritizing stylistic credibility.
• How Software Engineering Students Use LLMs to Write Research Papers: An Experience Report — Students use LLMs across academic writing, balancing utility for clarity and organization with critical verification of AI-generated content.
• False claims in a widely-cited paper. No corrections. No consequences. Welcome to the Business School. — Institutional failures enable influential, flawed research to persist uncorrected, undermining scientific integrity.
• i tested basically every AI tool i could find for med school research. most are useless unless you already know enough to catch them lying. — AI tools for professional research require extensive manual verification due to persistent accuracy and citation issues.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Nature Machine Intelligence.