id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
fpq-14292	Deery, Oisín; Bailey, Katherine	The Bias Dilemma: The Ethics of Algorithmic Bias in Natural-Language Processing	2022	29	.pdf	application/pdf	12976	674	55	In brief, even if we could eliminate bias from language models or their outputs, we would thereby often withhold descriptively or ethically useful information, despite avoiding perpetuating or amplifying bias. We must ask whether (1) it is practically speaking possible to debias language models and (2) whether we should do so, if possible.	cache/fpq-14292.pdf	txt/fpq-14292.txt