Discussion
You are a member of a grant allocation committee, composed of people that you don't collaborate with otherwise. You are expected to comment on another member's report on a project, which is very verbose and repetitive, suggesting the use of a chatbot. The committee has no AI policy.
What do you do?
@khinsen reject the proposal for its low quality alongside suggesting a policy against slow
@gedankenstuecke It's not the proposal that's slopified, it's a committee member's report on the proposal.
@khinsen ah, I misunderstood! Then definitely discuss with the colleague and if unsuccessful bring up to the whole committee!
@khinsen but then, there's a reason I don't have a traditional academic job - so take my suggestion with a grain of salt 😂
@gedankenstuecke The only reason why *I* have a traditional academic job is that I am old enough to have gotten that job when academia was still a more reasonable place!
@khinsen yep, I was born slightly too late for the more comfortable academic life!
@khinsen Initially I picked the option "Propose the adoption of an AI policy" having misread the premise as "The committee has a 'no AI' policy" And was put in mind of the Winston Churchill anecdote regarding a Silver cruet set. https://www.changingminds.org/analysis/churchill_salt.htm
Odds are this is a made up or misattributed as that's the thing about Churchill stories. However I thought this options would be a good way to allow a person to admit that they too were unaware there was a "no AI" policy. Of course if there is not already any agreement on the issue of the day then it's time to sound out colleagues on whether or not there should be. So in the end voted the same way but for less fanciful reasons.
@khinsen I would put that into the comment:
- it is verbose and repetitive
- suggesting the use of an LLM
- putting into question whether the member actually examined the project
If the comment is to be made synchronously (i. e. in a meeting), I'd probably hold back the last two points and first ask very specific and detailed questions about the project.
Make it obvious when someone actually didn't do their homework.
For the immediate situation, I think I would just comment on the report, pointing out that it is verbose and repetitive.
I would also propose the adoption of an AI policy. It's not entirely clear from the question, but I am assuming this report is evaluating a proposal. This report can now be produced without reading the proposal or even one's own report. So you need a policy to make sure the actual work of evaluation has been done. 1/2
The policy should probably allow LLM use for polishing grammar. Maybe have an expectation that reviewers keep their working drafts.
If the colleague continues to produce wordy, repetitive reports, and they don't seem to be actually doing the work of evaluation, you could also recommend they not be a committee member in the future.
This is a good example of the emotional labor LLM use is requiring of us.
2/2
Install bonfire.cafe
Get the full app experience
