Read on Twitter
So, brief story. Before my latest medical leave, I was working with very brainy people on the topic of IA (among other things ; it's complicated, but for now let's just say just IA). See those people don't code, they DECIDE what public agencies use. https://twitter.com/curiosarama/status/1051012656759562248
I'm much less brainy if at all : I'm just an assistant. So part of my job was to do a weekly press review of every article that talks about IA, so they can be informed about the latest of what's going on in the IA world, because it's obviously going very fast.
Now. My bosses are/were good people to work with. They're nice people, on the top of being smart motherf*ckers.
But they all, and I mean all of them, white cis men over 40. With a very specific education on the top of that, and the social status that goes with it.
But they all, and I mean all of them, white cis men over 40. With a very specific education on the top of that, and the social status that goes with it.
So I'm doing that press review about whatever people are talking about related to IA. In particular they're interested, of course, about longer reads, because we're not playing here. We don't want the basic "what will IA be able to do in the future" types of articles for noobs.
And regularly, articles come up with an angle that 1/ talks about the responsability of IA 2/ speaks to me. So I give it the old try and include ONE such article in my press review.
It's usually an article such as the one I quoted earlier. There are many like it.
It's usually an article such as the one I quoted earlier. There are many like it.
That bias, see, is not new. And the worrying part is, that bias should be of interest to my bosses, because they are the policy-makers that decide if IA is relevant to *public services*. That bias needs to be addressed, at the very least.
But every one of my attempts at including (not making the press review all about that theme, see, just INCLUDING) such an article, whoever wrote it, whoever published it, however long it is, however scientific... *has always been dismissed as irrelevant*.
See, the problem is not just that the algorithms have a bias at conception. It's that the people who have the power to decide whether to implement those algorithms or not have the same bias... and they want to keep that bias.
Please RT responsibly, when my medical leave is over I'd like to still have a job XD
I RTed the tweet below but then i was reminded of this thread ^ https://twitter.com/krazykitty/status/1307598884202598402
"Forgot".
