Read on Twitter
The New Yorker has a big piece on AI ethics out today (quoting lots of great folks) -- but with an overall framing that doesn't sit well with me, and particular characterizations that make my warning lights blink
Like this one, for instance.
OK, to elaborate: I'm of course thrilled when AI ethics works gets a high profile write-up, and thrilled to see scholars like Alex Hanna, Katie Shilton, and Brent Hecht quoted. And every long-form writer has their own style and own vibe, so fine. But: 1/n
As @KendraSerra notes, the framing of the piece is dismissive of CS/HCI critiques of binary gender classification as themselves sound science -- it doesn't dig into the nuanced work in the space by folks like @morganklauss, and worse... 2/n
it elides or ignores methodological and epistemological critiques of work like Speech2Face (for instance, through statements like this one: "Others argued, incorrectly, that a voice revealed nothing about its speaker’s appearance.") 3/n
If you're going to make a statement like that, you need to back it up! 4/n
More broadly the piece engages in a one-sided way, with the Twitter conversation around AI ethics. It's notable for the things it doesn't mention: the firing of Timnit from Google and the travails of the rest of the AI ethics team (including Alex); and the abuse... 5/n
and harassment of many AI ethics scholars On Here, especially women (and ESPECIALLY Black women). In fact, it even features praise, anonymized, for one of the worst offenders!). 6/n
Worse, it takes a few people's misreading of a study headline as evidence that those concerned about these issues are "trigger-happy with their moral outrage." We also get a wink at critics of "Cancel culture" ("it’s likely that more would have spoken up but feared backlash").7/n
Instead of broader engagement with critics, the piece caricatures them. Those humourless overly emotional woke scolds -- they're at it again! 8/n
As @Sharronapearl points out, the piece ignores STS perspectives on AI and computing more broadly. Exhibit A: referring to the NeurIPS broader impact statement process, the author writes that, "Some people on Twitter protested the intrusion of ideology into engineering." 9/n
That statement reads to me like a descriptive one (though if it wasn't meant as such I'm happy to be corrected) -- but makes it sound as if the default opinion of the reader *should* be that engineering is value neutral. 10/n
And finally, don't get me started on how subtly derogatory it is to describe Alex Hanna, a Google research scientist and quantitative sociologist as an "educator" in contrast to the "real" scientist responsible for Speech2Face. 
11/n
11/n
