# ICLR lets researchers use LLMs to analyse results but makes the authors responsible for verifying them, and requires a human author even for research an LLM produced

Page: https://flyvolo.ai/en/changes/ev-20250826-ai-researcher-6
Date: 2025-08-26 · Stage: Constraint
Can this record move a task judgement? yes
Occupation: https://flyvolo.ai/en/careers/ai-researcher

## Scope

One of the main machine-learning conferences, binding on its 2026 authors. The program chairs say LLMs can help with research ideas, experiment code and analysing results; that it is the author's responsibility to verify and validate any research contributions made by an LLM; that even where an LLM might produce an entire piece of research a human author is still required for accountability; and that researchers must not misrepresent results. Violations can lead to desk rejection. It allows machine analysis; it keeps answering for what the result means with a person.

## Source

- ICLR 2026 Program Chairs — 'Policies on Large Language Model Usage at ICLR 2026', ICLR Blog (26 August 2025) — https://blog.iclr.cc/2025/08/26/policies-on-large-language-model-usage-at-iclr-2026/ (primary source)

## What this means

In machine-learning research, the conference rules let the model help analyse a result but leave the claim about what it shows to a named person.

## What it does not show yet

One conference's rules for its authors; they do not measure how much analysis researchers now hand to models, or how well authors check it.

## How to verify it yourself

Open the ICLR Blog post 'Policies on Large Language Model Usage at ICLR 2026' and find "we still require a human author for accountability".


More: https://flyvolo.ai/llms.txt