AI usage policies

Our lab’s AI usage polices are based on a set of guiding principles:

• Thinking is an esssential aspect of writing. If we offload our writing we thus offload our thinking, which is fundamentally in conflict with our role as scholars.
• AI tools are meant to supplement and accelerate human expertise, not replace it.
• A human must ultimately understand and take responsibility for all outputs that go into a publication.
• AI usage for any aspect of a research project is a first-class decision that should be discussed beforehand.
• Because many journals require disclosure of AI usage, it is essential to record all of the ways in which AI is used in research project.

AI tools may be used for any part of the research process except for the drafting of narrative text for manuscripts. Lab members are expected to manually draft all of the text that they put into a paper; no copying and pasting of text from an LLM into a manuscript is allowed with explicit discussion and agreement.

It is the researcher’s ultimate responsiblity to vouch for any AI-generated outputs, so those outputs should be validated and verified just as one would for an output from an untrusted research assistant.

For all collaborations (internal and external), the policies for AI usage should be discussed prior to undertaking any work.

It is also the researcher’s responsibility to disclose to collaborators which outputs relied upon AI tools, and also to outline the steps that were taken to validate the AI-generated outputs. For any output, it needs to be clear what human is willing to stand by the output.

LLMs should be used to review all manuscripts prior to journal submission, preferably with an admonition that the LLM should not rewrite the text; rather, the author should rewrite relevant sections manually based on LLM feedback.

We expect all collaborators (internal and external) to adhere to this policy; any deviations must be explicitly discussed prior to the collaboration.