I strongly feel that this is an insult to life itself
ChatGPT has thrust Large Language Models (LLMs) into the public consciousness over the past three months. I’ve been writing about this for years, so I hope that lends some credence to what I’ll say here.
around the ethics of torturing simulated agents (see, e.g., Darling 2016). We are not aware of any laws or Institutional Review Board policies prohibiting mistreating simulated agents, at this time.
But this paper is not alone. There are at least three working papers using LLMs to replace human subjects in social science research. Miyazaki summarizes my response:
Replacing humans is bad. We study humans! How could we possibly want to eliminate our object of study — to eliminate humans?
LLMs represent a radical acceleration in the capacity of the production of text and thus of our informational crisis. We need to start figuring out how to adapt to this technology today—but I don’t think we currently have that capacity, mired in crisis as we already are.
This is among the most important normative questions of our time. The vast majority are underestimating the revolutionary potential of LLMs and related AI technology. So now is the time to have this debate, for decisive action.
