[Nettime-fr] "Le populisme algorithmique (Facebook)"

[Louise Desrenards <louise.desrenards@free.fr>]

Il faudrait traduire cette brillante analyse de Benjamin Bratton
envoyée hier sur sa page facebook de groupe. Mais je suis épuisée ; si
quelqu'un s'en ressentait l'énergie ce serait un joli cadeau à tous...

Après une série de questions réponses échangées avec lui entre deux
niveaux d'information -- lui en haut moi en bas -- dans la nuit, sur
la page d'un ami, en dépit de ma grande difficulté de m'exprimer en
anglais et donc de transmettant pas vraiment ce qu'en réalité je
voulais dire mais quelque chose d'approché, (beaucoup de risques de se
retrouver humiliée mais vous connaissez the fox  ça ne l'empêche pas
de foncer), finalement tout s'est bien passé parce que ce monsieur est
vraiment courtois... :-)

Voici en tous cas une brève recension de ma critique -- si j'ose dire
-- de son texte, qui m'a à la fois passionnée mais qu'en même temps
j'eus envie de le repousser et chacun peut comprendre pourquoi, étant
donné ce qui vient de se produire là-bas et se produira peut-être ici,
après que tout ce qui reste de la gauche en miettes s'y soit acharné
même si les miettes appellent à aller voter aux primaires de la droite
ce dont je me garderai bien.

Comme a dit notre "fort en gueule" national : "Ils voulaient voter
Sanders, on leur a dit de voter Clinton, ils se sont retrouvés avec
Trump"...

Si je peux être d'accord avec l'analyse sur Facebook et son système
sémiotique d'intégration de la société en cours, qui en renvoie une
image surfaite, une autre image, algorithmiquement sélective qui se
construit, de fait une image éditoriale qui à travers l'administration
obsolescente des gestionnaires de Facebook plastiquement s'énonce...
par contre je ne suis ABSOLUMENT pas d'accord avec la conclusion qui
ressaisit l'analyse dans l'avant dernier chapitre.

Mais oui il est certain que nous ayons besoin d'autres plateformes
d'échange et de partage que Facebook étant donné à l'évidence que
Facebook à travers ses filtres son administration des algorithmes
qu'il génère, des annonces de la pub des opinions cochées et des liens
partagés, est une façade éditoriale -- algorithmique elle-même
finalement conductible, même si elle ne s'affiche pas comme une
politique. De fait ses protocoles dynamiques l'élaborent à
l'entendement dans la direction des résultats stochastiques qui
sélectionnent les images instruites au premier plan, qui créent un
appel commun. Ce qu'il appelle le "populisme facebook".

D'abord je pense qu'il ne faut pas oublier de quitter le système
représentatif en miroir -- si dynamique soit-il ce qui ne ferait que
le rendre pire -- pour évaluer l'ensemble du dispositif électoral
appelé populisme.

Ensuite, le concept actuel de populisme (celui de Chantal Mouffe) ne
constitue pas une ligne éditoriale ni une ligne politique et donc
surtout pas le leader qui s'en sert, mais l'émergence des masses
post-démocratiques destructurées réagissant à des personnalités sans
valeur de croyance, s'en dégagerait-il comme leurs leaders, mises sous
surveillance par cette fluctuation même de la masse. [Ici je
reviendrai peut-être]

Et là je ferai hurler l'auteur de ce long billet qui médite à la lueur
des résultats électoraux aux USA sur une conférence de Mark
Zuckerberg, où lui-même se trouvait invité à un autre propos, si j'ai
bien compris...

Mais la plateforme des Insoumis pour soutenir à la fois financièrement
et électivement Mélenchon, et constituant une stratégie qui s'invente
dans le protocole participatif en désertant l'opinion, pour ne pas
être empêchée d'avancer d'un côté, avec des forums alliés où l'opinion
s'exprime d'un autre côté, teste peut-être un exemple algorithmique de
ceux que l'auteur de la longue note suivante préconise, ou semble
prédire, ou dont il avertit, un nouveau réseau social d'un autre type
peut-être, une nouveauté définie par des objectifs a priori annoncés,
et l'expérimentation aléatoire du partage de cette stratégie pour y
parvenir, mais qui s'invente en cours de route. Et en effet ici : un
populisme que l'on pourrait qualifier d'ouvert.

Mais cela justement les partis en France le rejettent car cette
aventure d'action directe troublant "leur ligne éditoriale" propre,
ils n'en veulent pas.

On peut toujours se rassurer -- ou se terroriser -- que Facebook ait
pu construire les résultats électoraux du scrutin qui découragent les
professionnels intellectuels et politiques américains, déshérités de
leur leadership avant-gardiste moderne, il y a encore peu crédibles
dans leurs prédictions des lendemains qui changent et/ou qui chantent
ou déchantent -- ou qui n'empireraient pas (mais ce serait ignorer
l'entropie), -- mais malheureusement rien n'est plus fort que le
refoulement du désir quand il se venge du pouvoir.

Qu'est- ce que le pouvoir pour "l'âme massive" et en partie
considérable changeante du populisme ?

Sinon le moment exact où, sauf la mort un citoyen, qu'il ait un toit
ou pas, rencontre l'absence de ressource, le coût du manger, les
flics, le juge, les taxes, et/ou le fisc, comme un empêchement
matériel de vivre (on voit déjà que cela recouvre plusieurs et
diverses couches asociales et sociales pouvant être bloquées de
vivre), dans le cadre de savoir que d'autres ne les rencontrent pas --
 c'est cela : le pouvoir.

Je garderai donc mon matérialisme vigilant...

L.D.


- - - - - - - - - - - - - -


Benjamin Bratton
Yesterday at 08:42
https://www.facebook.com/permalink.php?story_fbid=1008695262609699&id=956066011205958

A quickly-written (long) note on FB and algorithmic populism (on FB):

I was in attendance at the event where Zuckerberg discussed NewsFeed's
impact/non-impact on the election and the Filter Bubble more
generally. It was the same event where I was debating Kurzweil with
Vivienne Ming about A.I.

Much of the discussion among the attendees was critical of his
responses. The sentiment toward a stronger editorial hand to ensure
credibility/ validity of source/content was clear. That is of course
easier said than done, and comes with punishing trade-offs, both
obvious and not.

As mentioned in the BI article linked below, Google's PageRank (and
other Search and Display algorithms) were designed to surface the most
credible source on topics, based originally on peer-review citation
mechanisms, and in theory FB can explore and implement something
similar. Doubtless FB has and is re-debating that turn now.

Recall, however, the pushback from the Right when it was suspected
that NewsFeed's editorial team was supposedly "tilting" against
pro-Trump links (some of which were surely FakeNews and some not).
Here the Left populist sentiment against FB (in US but also stronger
in Europe) was, in practical terms, aligned with the Right's stance
that the Silicon Valley was using algorithmic mind control again.

Here some of the dilemmas of algorithmic populism may be identified:

— Algorithmic populism is and is not an opposite of algorithmic
autocracy. The former may undermine the latter, or enable it, and vice
versa.

— Zuckerberg is now in the awkward position of having to convince
people that FB does influence purchasing decisions but does not
influence voting decisions. How so?

—As discussed with @profcarroll this morning, Google’s incentive to
“protect integrity of intent-based ad biz” is clear, but Facebook’s
"incentive to throttle attention-based ad rev” is not. Fair point. But
as intention vs. attention oscillates, and as the latter bends toward
the former, do the incentives as well?

—Platforms are not just “media” they are governing systems, for ill
and good, not only in theory but in fact. An equally important
question as how FB “effects” politics is how FB displaces, absorbs or
becomes a polity. See Cloud Polis in Cloud Layer.

—In the interview on Thursday, Zuckerberg was asked by Kirkpatrick
about an episode in which FB editorial asked him to sign off on
categorizing some of Trump’s remarks as community terms violations, as
“hate speech.” He intimated that if the remarks were not from the
nominee of a major party, and thereby “mainstream” political
discourse, he may have done so, but would not do so now. He chose not
to decide that these remarks were exceptional but rather “normal”
enough that norms may continue. The scenario of information
infrastructures governed by fulcrum sovereigns is not a Speculative
Design scenario; we have that. He was, to use the parlance of another
time, asked if there was a "state of exception" and to decide on
suspension of norm/ or not and he chose “not." But as the election
makes clear, polities based on legal deliberation of “voice” and
Westhaplian jurisdictionally are also vulnerable to the pitfalls of
“deciding” bound up in one office, one person, one temperament.

—Zuckerberg’s response to continued prodding on how FB should
structure the community terms, the surfacing algorithms, and so on,
were (1) that FB tries to be as responsive as it possibly can to what
constituencies say they want and demonstrate they want by their actual
behavior, and I do not not believe him for the most part (but aside
from multiconstituency dissensus, that is also the crux of my strong
suspicions about algorithmic populism). (2) That NewsFeed is
empirically much more diverse that people seem to think (and much more
than old media), he says, and that that the problem is not that FB
fails to provide diverse information options, but rather that people
click on what already serves their conclusions, reinforcing them. He
is essentially shifting the Filter Bubble problem from platform to
user, from center to edge. Readers of The Stack will note that this
is, however, more an Interface Layer problem and will need to be
solved there. All interfaces are strongly reductive diagrams of
possible options (like all maps) and so the question of strong
editorial governance within the interfacial-image should be presumed.

—Zuckerberg was not asked about Peter Thiel’s role in the transition
team, which had been announced that morning. The two have had a
sometimes contentious relationship and I would imagine that Zuckerberg
would prefer to not ride on the long coattails of Thiel’s national
villainy. I don’t know. Maybe Zuckerberg means what he says and
honestly thinks that the longer term work of building global social
platforms will play itself out and the short term noise of elections
can’t stop the music. It would however be a mistake to conflate Thiel
and FB. There are plenty of other things to conflate him with at hand,
however. Among them is the role of scapegoating in the forming
national/ social mythologies. Thiel was a devoted student of
Philosopher, Rene Girard, at Stanford, whose important work on this
topic greatly influenced the young Thiel. Now that Thiel will help
install the scapegoater-in-chief, the weirdness comes full circle.

More general thoughts on algorithmic populism extrapolated from the above…..

At this moment it may be awkward to say to those who demand for more
and better Populism that they are wrong. But they are wrong. We need
platforms that structure what they mediate toward ends that don’t just
seem/feel better but actually are better. We need platforms that have
intelligent organizational outcomes. This is what infrastructure is
for. Easier said,….

— I think there is a deeper lesson than FakeNews (let’s also blame The
Onion?) regarding the Facebook Cloud Polis and the election, and this
the centrality of self-identity and self-expression as the binding
currencies of participation and credibility. As opposed to Amazon
which is based on the circulation of objects, Facebook is based on the
circulation of poses, claims, gestures and stances, and increasingly,
so too is what counts as politics. The platform amplifies their value
for the platform but also for its users.

— Populism is not (finally) kind of political platform. What may look
like Populism in Sanders’ platform is Center-Left technocracy in
Western Europe (it is not his fault that some of his supporters cast
him as a demagogue). The connotation of Populism to which I refer is
another kind of “aestheticiztion of politics.” It is a politics of
expression, of tribalism over facts, of gesture more than structures,
relationality and experience over realism, in which these are tuned to
a simple vertical image of ‘us vs them’ mapped into ‘low vs high’ and
‘authentic vs abstract’ as much as ‘in vs. out’. It follows to and
from the “theological” image of Politics that Carl Schmitt, Cornel
West, Ted Cruz, and Judith Butler have argued, quite variously, is
essential. It absorbs thought into stupidity, such as Franco Bifo
Berardi’s endorsement of Beppe Grillo’s Five Star movement (has he
recanted this?) Across the two (a “theological” aetheticization of
politics plus the absorptions of stupidity) we have the symptomatic
dreamworlds of conspiracy theory (another post, another day).

—There are some who this week will march in the streets to protest
outcomes, but who would not vote last week because doing so would not
have been an unbefitting gesture. It would have been the performance
of an unsuitable voice or identity perhaps. Together, this coalition
of the unwilling may have swung the election. Third party voters’
vision of politics as a theater of spiritualized self-expression is
known, but those who did not show up (for some because the opportunity
to make a sufficient act of passion was not provided by the brands on
offer) together proved to be the decisive constituency.

—I am not assigning “blame” for the outcome to them. Blame is not the
same as cause and effect, and it belongs to the very politics of moral
display that I am going on about. The world we share is automated by
systems laid down over the last years and centuries that we
continuously replace and displace. Identity may be performed into
existence, gesture by gesture, but reality is not. This is something
46.9% of eligible citizen voters in middle North America didn’t
discern, and the rest of the world is now on the hook for it.

—Platforms are not reducible to states or markets. And so, if
platforms like Facebook are going to continue to do, serve, or mediate
the work of governance (above described) then these platforms in turn
must be governed, and they are…. but…. Perhaps focus groups and
councils of sovereigns working to ensure maximal populist/behaviorist
death by a thousand clicks (as Facebook currently has) is a too
primitive model by a long measure. That said, simple majoritarianism
is unlikely ensure the best ends either: an infrastructure for
deliberation is not the same as an infrastructure deliberated (unless
the prospect of getting case-by-case community buy-in every time you
turn on the faucet seems like a good idea to you).

—Governance is a name for the ongoing design and designation of that
replacement and displacement of inherited systems. It requires
deliberation, and also sometimes tectonic shifts to and from
punctuated equilibria. Somethings the former causes the latter,
sometimes the latter stages the former and sometimes they are running
at different rhythms. Today I really (really) hope that they are.

In another post, I would be due to follow-up the criticism of
algorithmic populism with another set of points on alternatives to it.
They would involve (1) stronger not weakened means for platforms to
automate the disclosure and disenchantment of the real and to act back
upon it, and (2) dissipated incentives for self-affirming, posturing
symbolism to carry the day instead. (3) It would include references to
The Man in The High Castle, why “telling it like it is” and “artisinal
authenticity” are two sides of the same reactionary coin, and once
again, that the rise of ethnonationalist populism is a global
phenomenon with global causes. The demographic/ constituency split is
parallel in each case, and in each case the local Left mourns its
shame and the local Right congratulates its luck. Both are overstating
their role. We are trying to deal with a global shift one 18th century
jurisdiction at a time, and we need other means. Privileges of Modern
citizenship have been filtered by race and gender filters from the
start. It cannot be the only platform of voice, exit, loyalty,
strategy, decision, etc. The future has not been cancelled.

TW : @bratton

http://www.bratton.info/

http://visarts.ucsd.edu/faculty/benjamin-bratton

http://www.e-flux.com/journal/53/59883/the-black-stack/





-- 




-----
_______________________________________________
Nettime-fr mailing list
http://www.nettime.org/cgi-bin/mailman/listinfo/nettime-fr