Please register for the seminar here.

Discussion By:

Prof Simon Chesterman

David Marshal Professor,

Vice Provost (Educational Innovation), NUS

William Gibson reportedly said, “the future is already here – it’s just not very evenly distributed.” With surveys showing widespread large language model (LLM) use in scholarly writing and peer review, this seems apt for academia.

Just two years ago, LLM ethics discussions focused on known issues like bias and hallucinations, debating if they should be used; now, different questions seem more relevant. With sakana.ai claiming the first fully Al-generated paper passing peer review, and LLM performance on complex tasks doubling every seven months, we must seriously consider the impacts as LLMS begin automating academic processes. Traditional conceptions of merit, evaluation, production, peer review, and grants may lose ecological validity. Those mastering these changes could gain significantly, while others risk being left behind, creating a digital divide unlike any seen before. How should we navigate this transition?

Drawing on our work on AUTOGENs, ensemble methods, proposed ethical guidelines, and credit/blame, this talk outlines current capacities and the state of ethical debate, while exploring deeper questions about the future of scientific progress and academic labor.