🇸🇪 EU bör inte stifta lagar som förenklar för stater med auktoritära ambitioner att hämta ut data om medborgare

Foto: Wesley Tingey (Unsplash)

Ny debattartikel (27 okt 2019)

Frågan om så kallad e-bevisföring och möjligheterna för polis och rättsväsende i enskilda länder att kunna hämta ut persondata om misstänkta i brottsutredningar från elektroniska tjänsteleverantörer (som till exempel Microsoft, Google, Facebook) är högaktuell. EU-kommissionen visade i en rapport från 2018 att mer än hälften av alla brottsutredningar innebär en gränsöverskridande begäran om tillgång till elektroniska bevis.

Det finns redan instrument för gränsöverskridande samarbete (så kallade ömsesidiga rättshjälpsfördrag (MLAT) och lagen om europeisk utredningsorder (EIO) – men många regeringar har börjat mena att dessa verktyg ofta är för långsamma och besvärliga. Mot bakgrund av detta beslöt USA under 2017 att gå vidare med ny lagstiftning som skulle tillåta att polis och rättsväsende kan begära ut data från internet- och molntjänstleverantörer på mer eller mindre unilateralt sätt, direkt till polis- och rättsväsende i det land som utfärdar ordern om uthämtning av data, men med rätt olika villkor beroende på vilken plats informationen lagras eller var den misstänkta i en brottsutredning är bosatt eller befinner sig.

EU har inte varit långsamma med att svara på detta. Efter att USA:s kongress i mars 2018 anammade den så kallade Cloud Act som möjliggör nya direkt tvångsuthämtning av data oavsett om den finns i eller utanför USA, så föreslog EU-kommissionen i april 2018 “E-Evidence”, ett eget lagstiftningspaket som i princip utgör det europeiska motsvarigheten till Cloud Act och på liknande sätt syftar till att effektivisera samarbetet med tjänsteleverantörer och förse brottsbekämpande och rättsliga myndigheter med snabba verktyg för att få e-bevis.

Men i december 2018 klubbade EU-ministerråd ett förslag på reglering som går längre än kommissionens, och dessa skärpningar är riktigt problematiska. Flera av rådets ledamöter protesterade, rader av människorättsorganisationer och även näringslivets representanter med, men rådet presenterade ändå sitt förslag, och det är också detta förslag som jag kritiserar i min aktuella debattartikel i SvD.

I rådets förslag förbjuds tjänsteleverantörer att självmant meddela sina användare om att deras uppgifter begärs ut. I likhet med Cloud Act hotas de drabbade företagen med vite om de ens avslöjar att en begäran har skett. Undersökningens sekretess prioriteras över den drabbade personens rättigheter: Rådet vill att endast den stat som åtalar får möjlighet att meddela den misstänkte, och att överklagan endast kan ske i denna stat – som ju kan ha ett helt annat språk och rättssystem än den stat där personen är bosatt.

Tänk om Ungern kriminaliserade kritik av det egna landet i sociala medier. En svensk medborgare kritiserar Ungern på Facebook. Skulle du vara bekväm med att ungersk polis skulle kunna begära ut information om personen från Facebook och att lagen tillät detta, ja rentav förbjöd Facebook att bestrida det ungerska beslutet? Och om personen i fråga ville bestrida beslutet skulle denne tvingas åka till Ungern och göra det i ungersk domstol.

Kanske du skulle föredra att Sverige skulle fortsätta ha möjlighet att skydda landets egna medborgare från sådana uthämtningsorder?

Scenariot ovan skulle bli verklighet om EU anammade den lagstiftning som europeiska ministerrådet nyligen lade fram.

EU-parlamentets utskott för medborgerliga fri- och rättigheter (LIBE) är ytterst skeptiska, och tros rösta ner rådets förslag under kommande förhandlingar (se deras officiella rapport här).

Gränsöverskridande brottsutredningar måste kunna ske – och måste effektiviseras – men detta kan och bör ske i samverkansavtal, utan överilade inskränkningar i medborgares rättssäkerhet, länders nationella suveränitet och företags möjligheter att leva upp till dessa principer.

Referenser / vidare underlag i frågan:

Summering i frågan av EDRi (European Digital Rights, en förening av människorättsorganisationer runt om i Europa)

“E-Evidence In A Nutshell” Theodore Christakis (Université Grenoble Alpes). Cross-Border Data Forum, 14 januari 2019.

Extern länk till min artikel (betalvägg). Kontakta mig för tillgång

🇬🇧 Platform Logic: An Interdisciplinary Approach to the Platform-Based Economy

In September 2016, I attended an excellent academic conference, The Platform Society, arranged by the great people at Oxford University’s Internet Institute. My paper, presented there, was later re-worked into this article.

The concept of platforms has emerged in recent years as one of the most important concepts of the digital economy. In brief, my article concludes that digital platforms enact different types of governance, by recourse to three levels of observation: micro, meso, and macro.

  • In the minute, discrete interactions between platforms and users, micro-level forms of technocratic control are enabled.
  • On the level of platform interoperability (the meso level), a range of generative outcomes are supported.
  • In global aggregate, a macro-level mode of geopolitical domination is enabled.

Over at Oxford University’s Policy and Internet blog, you can read an interview with me about the article.

What’s the background to this article?

Digital platforms are not just software-based media, they are governing systems that control, interact, and accumulate. As surfaces on which social action takes place, digital platforms mediate — and to a considerable extent, dictate — economic relationships and social action. By automating market exchanges they solidify relationships into material infrastructure, lend a degree of immutability and traceability to engagements, and render what previously would have been informal exchanges into much more formalized rules.

Platforms enable a great number of new, seemingly rational and efficacious ways of organising society; but they are also based on an element of control, since users’ latitude is circumscribed by the computer code, and users are in many ways forced to adapt their behaviour to the interactions allowed for and prescribed by the platform owners.

A few platform-based corporations (Google, Facebook, Apple, Amazon, Microsoft) have gained massive global influence, since not only users but also a long list of other societal actors have become dependent on the services provided by these global companies, including many smaller, upcoming platform companies.

How does my concept of “platform logic” become useful?

If one chooses to look at the discrete, often highly technical inter-platform affordances and connections, one will see generativity and scope for innovation. This is what is often focused on in the business press, and similar outlets, despite the fact that many economists would argue that our present era of digital development is less innovative than past ones.

If one chooses, instead, to look at the emerging transnational, geopolitical formations under platform capitalism, one will make an entirely different set of observations. Theorists like Nick Srnicek and Frank Pasquale have argued that platform capitalism begets historically unprecedented forms of economic domination.

Lastly, if one chooses to observe the very minutiae of platform interaction — the ways in which individuals and organisations adapt to the technical imperatives that the platforms as infrastructures implement, one will see that there is a strong form of technocracy involved. Researchers like Robyn Caplan and danah boyd have recently shown how this takes place in institutions, as different organisations adapt their ways of doing things so that they become more compatible with the existing platforms, and in order to emulate the alleged efficacy and agility of tech companies. I, myself, have argued that the epistemological convictions that are at the root of behavioural data-gathering companies such as Facebook, and the technical prescription exerted by the resulting infrastructures, might be much more rigid than many would think, steering also the operatives inside the platform corporations to an extent that we should not underestimate.

The interplay between these different mechanics (each one observable by using the attendant optic) can be neatly summarized by my concept of “platform logic.”

I argue that platform logic is both conforming to and distinct from pre-existing capitalist structural logics (Taylorism perhaps being the one closest at hand, something that was recently seized upon by Evgeny Morozov in his long review of Shoshana Zuboff’s Surveillance Capitalism). Due to the digital nature of platforms, many tendencies already latent in capitalism (monopolism, colonialism, generativity) are exacerbated, while some altogether new tendencies can also be observed.

Platform power can be summarised as ‘the power to link facially separate markets and/or to constrain participation in markets by using technical protocols’ (Cohen 2016: 374). Data is generated, almost automatically, the very moment the infrastructure is used, enabling surveillance and various designs that utilize such data. This has primarily been discussed in relation to the distribution of ads and editorial content in the media sector, but has huge importance also for other industries. Further, digital platforms directly benefit from so-called network effects that make the platform exponentially more valuable as more people use it.

We already know that digital systems have the quality of being possible to scale, virtually endlessly. We also know that code is control, in the sense that events aboard platforms can be governed in absolute, binary ways; users and possibilities can be turned on or off. However, this hard logic of infrastructural control stands in tension with the softer, more generative potentials that are often observed as inherent to digitization; programmability, interoperability and so on. In other words, platforms are charged with a ‘paradoxical tension between the logic of generative and democratic innovations and the logic of infrastructural control’ (Eaton et al. 2015: 218). My concept of “platform logic” refers to this quite specific and, at times, paradoxical interplay that platform power results in.

Andersson Schwarz, J. (2017). Platform Logic: An Interdisciplinary Approach to the Platform-Based Economy. Policy & Internet, 9(4): 374–394. DOI: 10.1002/poi3.159

Paywalled. Contact me for access

🇸🇪 Plattformssamhället

Digitala plattformar har under de senaste åren seglat upp som ett av de mest centrala begreppen i den digitala ekonomin. Plattformar möjliggör mängder av nya, effektiva sätt att organisera samhället – men de bygger också på ett element av styrning, då mänskligt handlande måste anpassa sig efter datorkoden. En handfull plattformsbaserade företag (Google, Facebook, Apple, Amazon, Microsoft) har fått enormt stort globalt inflytande, där inte bara användarna utan rader av andra samhällsaktörer har blivit beroende av tjänster från dessa giganter.

Samtidigt har de senaste åren rader av mindre företag dykt upp (många av dem nyetablerade så kallade startups, finansierade av riskkapital), vars affärsmodeller är baserade på olika typer av plattformar. Även många av dessa mindre plattformsföretag är i många avseenden beroende av dem.

Tillsammans med Stefan Larsson satte jag under 2018 ihop en antologi med syfte att samla flera kloka svenska röster i frågor relaterade till plattformsproblematiken. Publiceringen av antologin Plattformssamhället (Fores) i början av 2019 sammanföll i tid med den internationellt inflytelserika boken The Platform Society (Oxford University Press) av de holländska ledande forskarna José van Dijck, Thomas Poell och Martijn de Waal, som i synnerhet lyfter fram plattformarnas snabba framträdande inom olika områden som i Europa länge har varit offentligfinansierad verksamhet medan i USA är utpräglat privatägda (urban transport, nyhetsproduktion och dissemination, hälsovård och utbildning). Under hösten 2018 stod jag som extern reviewer av holländarnas bok, och arbetet med den svenska plattformsboken gynnades av denna internationella utblick.

Digitala plattformar kan få progressiva och rentav livsavgörande (goda) effekter – men kan likaledes användas för att kontrollera, manipulera och övervaka människor. De stora plattformsföretagen har global räckvidd, men det finns skäl att förutsätta att jurisdiktioner och politiska system förblir nationella. Häri ligger en rad utmaningar. Flera av dem berörs i boken.

Andersson Schwarz, J. & S. Larsson (red., 2019). Plattformssamhället: Den digitala utvecklingens politik, innovation och reglering. Stockholm: Fores.

Länk

🇬🇧 Umwelt and individuation: Digital signals and technical being

This chapter, which forms part of a deep and existentially far-reaching anthology on Digital Existence, is essentially a plea for a more responsive, cooperative information infrastructure. I address this by taking Facebook as an example.

Today’s digital landscape is quite literally premised on a theory of information that was in fact intended for machines – Claude Shannon’s theorem from 1948. Thus, the digital imaginary of our time is unfortunately of a very rigid, mute, non-vitalist kind – essentially inhuman.

My chapter is an attempt at reaching towards a more integrated, dynamic, vitalistic, and inclusive theory of digital information, by adopting the theory of Gilbert Simondon, a French 1950s thinker of technology.

Simondon affirms technology as a symbiotic process, enabling a utopian future where humans and digital infrastructures can be allowed to truly co-habit this planet – in contrast with today’s mainstream paradigm, which rather seems to stipulate an alienated relationship to technology, humans in one ringside and machines in the other. In Simondon’s theory, the individual is not a being but an act, and individuality is always an aspect of generation, ever-evolving, an ongoing genesis.

This stands in stark contrast to prevailing technocratic “solutions” (apps, platforms, databases) that are essentially systems of control, where users are deprived of genuine participation and are at best offered limited forms of co-creation that are always conditional on the proprietors or owners in question. At worst, the participation allowed for users is only illusory. The very act of trying to encapsulate human being into predefined, finite and locked-down boxes – trying to “pin down” individuals and groups by recourse to palimpsests, intended to “freeze” system states as if these were reliable and objective snapshots of human behaviour – is reductive and regressive at its core.

Believe it or not: These rather outlandish epistemological convictions actually lay at the root of today’s tech companies that base their business models on behavioural data, leading the operatives inside of these companies to pretend that the signals gathered are truthful and representative renditions of human behaviour.

What is more, once these operatives implement new applications based on the data that they are constantly gathering and feeding into algorithmic systems of behavioural manipulation and control, these systems actually begin to actively shape the real world that they are interacting with.

Soon, sinister feedback loops emerge: By observing the behaviours that these algorithmic systems prescribe, indeed dictate, users are taught to behave in specific ways in order to navigate the interface in the expected ways. By doing so, they become enticed to make further interactions which will, in turn, be farmed into new, interesting content for other users to interact with: Think of how Facebook users are compelled to publish and share content that is expected to be desirable among their peers.

More importantly, any move that a user would make is monitored and recorded so as to enable the corporation to interpret these signals in order to make selections of content and advertisements that they believe that the user him- or herself would find interesting, based on what they read these signals to indicate.

Moreover, users would arguably adapt also their own behaviours in order to suit the algorithmic infrastructure: In order to maintain peer visibility, users are compelled to design their posts in accordance with what the algorithmic interface tends to value as popular or recognizable to a large audience (Gillespie, 2014: 183). This precipitates a kind of built-in conformism; a popularity bias (Webster 2014).

Algorithms indirectly construct culture by way of feedback loops like this. Individuals seem to act based on what they observe that these semi-automated systems seem to value.

My argument, in brief

There is a funny thing though.

Do you see how the humans in the loop always have to second-guess what the system would prefer or predict? Essentially, the corporation makes educated guesses from all the vast amounts of user signals that they collect, and try to make target groups and so that they can increase the chances for advertisers to place ads that actually engage the users. Essentially, users themselves try to “game” the system so that they can reap as many benefits as they can from using it.

Researchers like Taina Bucher and John Cheney-Lippold have come to similar conclusions.

In order to understand all of this better, let us think of these media-technological systems as Umwelts for individuals to roam through. The concept of Umwelt was developed in the early 20th century by the Baltic German biologist Jakob von Uexküll, and refers to the cognized environment, the “self-centered world” which all organisms live in. All organisms experience life in terms of subjective reference frames: a bumblebee is at the center of its own world, much like the Facebook user is at the center of her own world, uniquely personalised for her, by Facebook the corporation™.

So, as users interact with environments-that-are-unique-to-them-and-only-them, they would at the same time give off signals as they keep interacting with this built environment. After all, this is an environment that is built on surveillance, all the way through. These signals are then instantly harvested by the platform proprietors and are read to be indicative of the assumed internal states of these individuals.

The really clever thing with this argument, though, is that we can think of also the platform infrastructure’s intelligence as a form of technical Umwelt unto itself!

Facebook doesn’t magically “know” you, as if we were dealing with some kind of sentient fairy-tale being, a Leviathan of some kind (although some critical scholars would definitely seem to want to frame it like that!) The platform operators and managers can actually only “see” that which takes place in the direct interactions, the actual “clicks” and measurable movements made. This is, quite literally, all that the automated systems have to go on. A system is a sum of inputs. It is by compiling signals, encoded in the form of “behavioural data,” that the engineers, behavioural scientists and marketing experts who build and maintain this infrastructure make their decisions.

Consequentially, we should not underestimate the degree to which the actual operatives inside the platform corporations are informed by estimations that risk being very reductive, if not even blind to a lot of aspects of human life.

A stunning addition!

After having finished this article in 2018, I was reminded of the concept of affordances, pioneered by cognitive psychologist J.J. Gibson in 1979. It is a bit embarrassing that his work hadn’t actually crossed my mind before. I’m schooled in a field somewhat indebted to continental philosophy and the Frankfurt school, so the work of an American mid-20th century psychologist hadn’t really cropped up on may radar.

But, conversely, Gibson himself had no reference to Umwelt either.

Andersson Schwarz, J. (2018). Umwelt and individuation: Digital signals and technical being. In: A. Lagerkvist (Ed.) Digital Existence. London & New York: Routledge. 61-80.

Paywalled / contact me for access

🇬🇧 Heuristics of the Algorithm

As the cultural and media industries have developed into 21st century forms, where large aggregates of personal information (behavioural data) is mined in order to find patterns of correlations so that individuals and target groups can be identified, me and my co-author Göran Bolin explore some of the foundational heuristics that businesses have to rely upon.

We begin by contrasting 20th century audience statistics with those of the 21st century. 20th century intelligence on mass media audiences was founded on representative statistical samples, analysed by statisticians at the market departments of media corporations.

In the 21st century, an age of pervasive and ubiquitous personal media (e.g. laptops, smartphones, credit cards/swipe cards and radio-frequency identification), techniques for aggregating user data build on large aggregates of information (Big Data) analysed by algorithms that transform data into commodities.

While the former technologies were built on socio-economic variables such as age, gender, ethnicity, education, media preferences (i.e. categories recognisable to media users and industry representatives alike), Big Data technologies register consumer choice, geographical position, web movement, and behavioural information in technologically complex ways that for most lay people are too abstract to appreciate the full consequences of.

The data mined for pattern recognition privileges relational rather than demographic qualities. We argue that the agency of interpretation at the bottom of market decisions within media companies nevertheless introduces a ‘heuristics of the algorithm’, where the data inevitably has to be translated into social categories.

In the paper we argue that although the promise of algorithmically generated data is often implemented in automated systems where human agency gets increasingly distanced from the data collected (it is our technological gadgets that are being surveyed, rather than us as social beings), one can observe a felt need among media users and among industry actors to ‘translate back’ the algorithmically produced relational statistics into ‘traditional’ social parameters. The tenacious social structures within the advertising industries work against the techno-economically driven tendencies within the Big Data economy.

Bolin, G. & J. Andersson Schwarz (2015). Heuristics of the Algorithm: Big Data, User Interpretation and Institutional Translation. Big Data & Society, 2(2): 1–12. DOI: 10.1177/2053951715608406

Link