Detailed, readable documentation (which is accessible directly from the
search interface)
Yup, totally agree. But, in real life, who reads the documentation anyway?
;)
--
deb tankersley
irc: debt
Product Manager, Discovery
Wikimedia Foundation
On Mon, Jul 10, 2017 at 11:21 AM, Jonathan Morgan <jmorgan(a)wikimedia.org>
wrote:
When I read this article, I wasn't struck that the
author was saying she
thought that technology "owed" her particular results.
I think the point she's making is that so much of our life now is mediated
by algorithms that make choices we may not understand, and that impacts how
we see the world in ways we can't easily anticipate or account for
(supporting quotes below). And the problem is subtler and more pervasive
than simply issues of "filter bubbles" and "fake news" that are
currently
garnering the biggest headlines.
This is part of a broader conversation that happening right now around
algorithmic transparency and "ethical AI". Lots and lots of big names are
weighing in on the topic[1][2][3][4][5][6].
I haven't see a whole lot of specific design guidance around how to
support transparency in the context of search yet, but I'd be interested in
hearing from others who have. Detailed, readable documentation (which is
accessible directly from the search interface) sounds like a pretty good
start :)
- Jonathan
*"I am still not accustomed to the drastic ways search algorithms can
direct people’s lives. We’re so used to Google’s suggested spellings and
the autocorrect of texting apps that we’ve stopped thinking too hard about
how we search or how we spell. If I tap out Chrissy but should have typed
Krissy, I implicitly believe that of course the opaque algorithms of
Facebook will intuit my intent. But we have no way of probing the limits of
the algorithms that govern our lives.""When we talk about the algorithms
that drive sites like Google and Facebook, we marvel at their cleverness in
serving us information, or we worry about the ways in which they exacerbate
bias—profiling people based on gross data trends, for example, to decide
who gets a loan and who doesn’t. But there is a complex web of algorithmic
life-shaping at work that we barely register. It’s not that I wish Facebook
treated its Cs and Ks alike. It’s that by not knowing the rules, we give up
some agency to mathematical calculations."*
1.
https://www.acm.org/binaries/content/assets/public-policy/2017_usacm_
statement_algorithms.pdf
2.
http://www.pcworld.com/article/2908372/the-ftc-is-
worried-about-algorithmic-transparency-and-you-should-be-too.html
3.
http://www.pewinternet.org/2017/02/08/theme-7-the-need-
grows-for-algorithmic-literacy-transparency-and-oversight/
4.
https://epic.org/algorithmic-transparency/
5.
https://www.technologyreview.com/s/603915/tech-giants-grapple-with-the-
ethical-concerns-raised-by-the-ai-boom/
6.
https://cyber.harvard.edu/research/ai
On Mon, Jul 10, 2017 at 9:04 AM, Trey Jones <tjones(a)wikimedia.org> wrote:
Thanks for sharing, Chris!
I found the article a bit frustrating. As a human interest story, it's
very touching that the sisters were able to reconnect despite family
problems that worked to keep them apart.
But from the technology side of things, blaming search algorithms seems
odd to me. I'm surprised that anyone would feel that technology owed them
particular results or specific capabilities—especially capabilities they
didn't even know they needed. That might actually be a useful insight into
our own users, though.
I'm also surprised the author didn't use anything other than search
engines and social media. I've had to track down a dozen or so people who
were out of touch for up to 20+ years, for a book project, and there are so
many resources out there! Even more if you are able to spend a few dollars
per person—which "book project people" did not warrant, but siblings would.
So, getting a bit more on-topic, how do we help people by not only
providing them with useful information, but also the tools and processes
that allow them to get the most from that information? It seems like
documentation works for very sophisticated users, but the rest have to
collectively and very unevenly accrete familiarity with tools over time;
learning/teaching processes seems even more daunting. I can't see a way to
accelerate that process, which is disheartening.
—Trey
Trey Jones
Software Engineer, Discovery
Wikimedia Foundation
On Mon, Jul 10, 2017 at 10:33 AM, Chris Koerner <ckoerner(a)wikimedia.org>
wrote:
Thanks to Erica Litrenta for sharing this with
me. I thought I'd share
if forward.
"It was because of the letter K that I found my younger sister, but for
14 years, it was also the letter K that kept us apart."
https://www.wired.com/story/search-algorithms-kept-me-from-m
y-sister-for-14-years
Yours,
Chris Koerner
Community Liaison
Wikimedia Foundation
_______________________________________________
discovery mailing list
discovery(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/discovery
_______________________________________________
discovery mailing list
discovery(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/discovery
--
Jonathan T. Morgan
Senior Design Researcher
Wikimedia Foundation
User:Jmorgan (WMF) <https://meta.wikimedia.org/wiki/User:Jmorgan_(WMF)>
_______________________________________________
discovery mailing list
discovery(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/discovery