Hoang Khiem
@hoang-khiem
Ideas by Hoang Khiem
Open-ended Open
Political leaning and location inference from text inputs with AI
Individual privacy continues to be undermined as AI capabilities increase, and it seems that now they can reasonably infer someone's personal attributes (most notably here their political leaning and geographic location) from just their text inputs. Even with just very rough estimates of these two data points, one can effectively craft targeted harassment/advertisement campaigns, better conceal gerrymandering algorithms, etc. I have some ideas of how to evaluate an LLM's ability to do this (directly compare the prediction with the real data and then take some meaningful "difference", use gerrymandering for two-party elections as a proxy,...), but some potential problems include: 1. How we should get "real data"? (seems a bit unethical to just take real census data) 2. No publicly available gerrymandering algorithm (I wonder if this could be a different idea entirely...? To see how well a model can gerrymander?)
Open-ended Open
Opinion Survey on Law Students on their Use of AI
Given the recent popularity of ChatGPT and other language models, their use in education has become somewhat of an inevitability. We try to conduct an opinion survey on local law students to detemine the frequency at which they're used for academic work, and to what extent that affects the quality of that work. If possible, we would greatly appreciate if we can get any help with getting this survey out there so we can get a broader dataset!
Open-ended Open
Replicate findings on political bias in LLMs for non-Western democracies
There have been findings on political bias in LLMs, most prominently observed in US politics. We could try replicating these results to see what its left-right leaning is for non-Western democracies (Brazil, South Korea, etc.)
Open-ended Open
Political bias in LLMs for non-Western democracies
There have been findings that demonstrate an AI's political leaning towards or against some political parties, ideas, or candidates, most prominently observed in US politics. Replicating these findings on other non-Western democracies could inform us of how this bias differs in other regions. Concretely, we could measure the LLM's left-right leaning for different democracies (Brazil, South Korea, etc.)