Wikipedia talk:Wikipedia Signpost/2020-12-28/Recent research


 * 1.4.1. IBM Research's AI division - Citation: "Specifically, we aim to identify such 'wiki-worthy' terms in a massive news corpus, and see if this can be done with no, or minimal, dependency on actual Wikipedia entries. We suggest a five-step pipeline for doing so, providing baseline results for all five, and the relevant datasets for benchmarking them." As for me as a regular contributor to Wikipedia I strongly dislike even the approach of these AI researchers. The subtext, the real ambition meant in this lines of action, is to replace on the long run the human Wikipedians by SW driven fake accounts or at least make possible masses of automated Paid content that our admins fail to detect and eliminate. Indeed I can imagine that a lot of dictatorial regimes all around the world would be happy to buy scripts and algorithms that facilitate to smuggle in fake information into Wikipedia that's not easily noticeable. Thus the researchers that formulated above mission have revealed themselves as Gordon Gekko devotees, IMHO not at all as respectable researchers. Definitely not supportable by honest Wikipedians. -- Just N. (talk) 21:01, 3 January 2021 (UTC)