-
swodinsky ow I mean, if there is the right number of microphones in the right positions you can get a very good idea of the person's location in a home over time. If they are doing that is an open question, but seems like a solid bet.Permalink On twitter.com
❤️ 1 Favorite
Mood +7 🙂
-
swodinsky ow Beyond that we know this is an experimental tech aimed at consumers in order to have them train it so that's creepy in the sense of a targeted experimentation group. But flip side is, if this becomes mainstream this group would likely be left out of initial training otherwise.Permalink On twitter.com
❤️ 1 Favorite
Mood 0
-
swodinsky ow It's very easy to imagine a world in which your smart speaker might treat you differently based on your persistent height relative to the speaker, or on types of voice quality, b/c it is trained on abled people w/ different expectations, and badly treat the disabled as a result.Permalink On twitter.com
❤️ 1 Favorite
Mood +1 🙂
-
swodinsky ow I mean it is a proactive solve which is weird to think about. (Like good they learned lessons from Microsoft forgetting to train facial recognition on black people b/c none are in the testing groups, etc... ) but...On twitter.com
❤️ 1 Favorite
Mood +8 🙂
-
swodinsky ow At the end of the day, whether Google thinks about it this way or not, this is specifically about running experimental technology on a subject group selected due to their status as 'different from standard tech user group/employees' & that part is sort of fked up.Permalink On twitter.com
❤️ 1 Favorite
Mood 0