AI Policies

The following outlines the CLeAR lab’s current AI policies. These policies are a work in progress and will almost surely evolve in response to our changing environment.

LLMs are research tools

Treat AI assistance as you would any other research method. Use AI tools in your research or do not use them: regardless, your work must be held to the same standards of rigor, transparency, and professional integrity either way.

When AI tools contribute substantively to your work, document which tools you used, how and for what purpose you have used them, what they contributed, and how you validated their output.

Our responsibilities as researchers

As researchers, we are responsible for every claim, argument, derivation, result, piece of code, figure, and sentence we present, regardless of how it was produced. You must understand and be able to defend AI-assisted work as thoroughly as any other work.

LLMs do not change what is good writing

Using AI tools to help polish your writing is acceptable. However, submitting AI-generated prose that you do not fully understand, endorse, and make your own is not acceptable. Generic or formulaic AI writing, imprecise exposition, and unnecessary jargon are unacceptable regardless of how they were produced.

Lab rules about AI usage

Concretely, here are the current set of lab rules surrounding the use of AI tools:

• Every technical artifact (code repository, paper draft, idea on slack, etc.) must be accompanied by an AI usage statement along the lines above.
• Misrepresenting or concealing substantive AI use, or presenting work one does not understand and/or cannot defend, breaches research integrity.
• Producing AI slop, however properly acknowledged, breaches readers’ and collaborators’ trust. Do not produce AI slop.
• If you feel you lack the necessary judgement or discipline to prevent AI tools from compromising your professional integrity, do not use them.
• If you violate these principles of AI usage, there will be one and only one warning.