https://hackerpaper.kaipfister.de/p/1912.06979/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMThttps://news.ycombinator.com/item?id=25576158urn:hackerpapers:mention:1912.06979:25576158Tue, 29 Dec 2020 23:04:53 GMT<p>&gt; As far as I know, those are for experiments that directly interact with humans as part of the experiment itself. Or use certain kinds of data from humans. For example, a machine learning project on medical data (even anonymized medical data) probably needs an IRB approval to gain access to, say, a dataset of cancer imaging. A facial recognition model doesn&#x27;t. What makes those fundamentally different (arguably the facial image is more identifying). &gt; However the incident in question is really about a different kind of thing - it&#x27;s asking researchers to speculate on the future ramifications…</p><p><a href="https://hackerpaper.kaipfister.de/p/1912.06979/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.HC