Carsten Agger on Wed, 29 Nov 2017 22:29:02 +0100 (CET)


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

<nettime> Algorithmic and pathological children's videos


These are some speculations that were started by this article,
"Something is wrong on the Internet"
(https://medium.com/@jamesbridle/something-is-wrong-on-the-internet-c39c471271d2)
about a phenomenon which I wasn't aware of and is apparently also quite
new, exploitative clock-bait videos for very small children.

These videos' modus operandi is that a parent hands a toddler a tablet
or a cell phone and finds them a YouTube video. If youTube is on
autoplay, a "similar" video is played after that, and so forth. So, if
many parents hand their toddlers an Internet device, videos that are
"similar" to other videos that parents would find for their toddlers
will receive many views. Like, if the video comes up when searching for
"nursery rhymes learning video". If you haven't experienced this before,
I suggest you try to enter that search and see what comes up. Click on
one of the videos and look at the suggestions.

Now click on one of *those* and look at the suggestions. It seems clear
that a toddler left alone with a tablet is more or less bound to be
sucked in by a feedback loop of eerily similar videos, meticulously
crafted to imitate thousands of other videos that happen to have
received astronomical numbers of views, often in the millions.

This video: https://www.youtube.com/watch?v=d7p0nSW0RI8 is typical  in
many ways - it's simple, strange, creepy in an eerie way, and
algorithmically situated to come up as "related videos" when a 2-3-year
old is binge-watching YouTube. It has 15 MILLION views.

"Wrong heads" and "learn colors" is one of the most popular tropes.
Right now, there are three million(!) of them on YouTube. And this video
alone has more than six million views.

Another trope, which we could call "finger family learning video" yields
7 million hits on YouTube. The top search result (this
video: https://www.youtube.com/watch?v=ivFefy28M10) has 82 MILLION views.

So how did 7 MILLION of one kind of video, three MILLION of another, and
MILLIONS more of other tropes, e.g. "bad baby learns colors" which has
about 8 million, all come into existence the last few years?

What the actual fuck is going on here? Who make them and how does it work?

Of course, this has been under some discussion since James Bridle's
article (linked above), among other things under the topic #elsagate:

While many of the videos already discussed are sickeningly repetitive,
with titles obviously designed to attract algorithmic matches rather
than human eyes ("Wrong Dress Frozen Elsa Sofia Talking Angela Hulk
Finger Family Learn Colors For Kids"), others are decidedly
inappropriate and not a little creepy, including urination, defecation,
pregnancy and strong sexual allusions, all in material very obviously
targeted at toddlers without adult supervision.

And that is part of Bridle's and other people's angle: Can
binge-watching these, surprisingly *very* popular and apparently
numbering in the hundreds if millions, of videos be harmful for children
- what effect can this algorithmic tour de force have on toddlers left
hour after hour to themselves?

However, I also have another suggestion - that this video industry and
its objective, being watched by toddlers, displays a very strange
feedback loop that actually us something about the toddler's minds. In a
way, this video industry is a kind of AI investigating the ways of
thinking that appeal to very small children, what they feel is funny,
what can preoccupy them ... like looking at colors, listening to
recognizable nursery rhymes, seeing favorite cartoon characters do funny
stuff, watch unboxings of fancy toys, on the innocent side - but also
tricking people, not being able to go to the toilet, peeing in
inappropriate places, etc.

So in a way, this video artist/toddler feedback loop is an AI
investigation of the psychology of very small children, where the AI is
implemented not just by computers, but by thousands (apparently) of
video production houses doing both cartoons and live action all trying
to find the exact sweet spot where the toddlers just want to keep
watching - by pressing the very buttons that appeal to children that
way. As a feedback loop always changes what it reproduces, this also
appears to be an example of what the Brazilian pshychologist Fabiane M.
Borges (disclaimer, a very good friend of mine) calls "hacking the
unconscious" - the videos is hacking the unconscious, to use the
psychoanalytical term, of millions of very small children, thus
displaying what goes on in them - but, also changing them and being
changed by them at the same time, all mediated by YouTube's algorithms
(but possibly not for a lot longer, YouTube appears to be cracking down
and maybe this phenomenon will soon be history).

This may be only a much more blatant and in-your-face version of the
feedback loop we have seen with adults and television for many years -
but in that case, I find it quite eye-opening. If anyone has the stomach
for a further analysis (I'd generally recommend people *not* to spend
too much time on these videos, if they value their peace of mind) I'm
open to hearing other takes.

Best
Carsten


#  distributed via <nettime>: no commercial use without permission
#  <nettime>  is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: http://mx.kein.org/mailman/listinfo/nettime-l
#  archive: http://www.nettime.org contact: nettime@kein.org
#  @nettime_bot tweets mail w/ sender unless #ANON is in Subject: