Pax Technica: morning panel

Notes from a CRASSH event by the Technology and Democracy project. Updated through the day. Following on from first session.

Bill Janeway — doing capitalism in the innovation economy. The 3 player game reloaded — markets, state, speculation/innovation. Digital revolution no longer needs sponsorship from the state, instead it’s attacking the state at all levels (local to national) and even undermining the processes that give state legitimacy. Market economy in extraordinary situation — more companies dying than starting, IPO market gone, speculators focussing on unicorn bubble and funding the build out of the digital economy when the world needs is response to climate change. Financial speculators don’t have a vehicle to mobilise capital to respond to this.

Poses question: does China have the model? China in the context of digitalisation — great firewall on one hand, massive state sponsored investment in AI and similar on the other hand — prospect that there may be a great transfer of leadership of the innovation economy from US to China (following earlier transition from UK to USA in 20th century).

John Naughton —digital tech has special affordances. Near zero marginal cost, network effects, no normal distributions just power law, tech lock in, capacity for infinite surveillance. First 4 point to winner take all outcomes…. 2 questions. (1)Standards — the power resides in the capacity to set standards that stick. (2)What business models will enable and drive this technology in the future?

On standards, it boils down to — who or what will set the standards for how this tech evolves? past examples: TCP/IP — a set of technical stakeholders, not govt, not companies, but IETF; amazing way to build something powerful. WWW / W3C is another — but had intrusion of corporate stakeholders and the exercise of that corp muscle; seemed constrained for a while but possibly not any more, with DRM in HTML5 suggests corporate interest is dominating over public interest. Older model — the microsoft model — they had 95% desktop market, by de facto adoption of a proprietary standard. So: how are these standards to be set, by whom? We have 3 internets — effectively — how will this play out? remember astonishing botnet attack on Krebs, when even lightswitches and webcams were the attack units, not what we normally think of as computers. Easy to make an IOT device — bolt some bits together — no concern for security — market failure as organisations have no vested interest in security. Bruce Schneier —this is the sort of market failure only govts can fix.

What businesses models can work? The only one we have is based on surveillance capitalism, the internet of people, a vast real time market. We see now the consequences, with weaponisation of social media and so on. YouTube is central to a lot of the alt-right and dysfunctional politically motivated stuff. Could disable a lot of the objectionable stuff by shutting down YouTube. Or Twitter. This is the result of user engagement being the only thing that matters. We have to find some other models.

Bill — book — Globalisation Paradox — political Trilemma — proposes you can have two out of three: national autonomy, representative govt and deep tech enabled integration of markets for goods, labor, services and capital. Treaty of Westphalia — had lots of tech like telegraph, steamship. Suggests some nuance for today.

Marginal cost = zero, suggests only two equilibrium positions. Winner takes all, OR regulated service provision (possibly locally regulated). Providing electricity, and rail transport, are both zero marginal cost too… Debate on these cases (rail etc) was where the frontier would be between these two positions. Note big populist rural revolt in USA when railways dominated the scene. Today, we have a particular challenge, different in different countries because of different national cultures. USA living with consequences of a generations worth of collaboration/conspiracy between intellectual entrepreneurs and enormously wealthy private sector actors, who delegitimised government as a source of power. USA had weird racial thing around 2009 which meant Obama govt actions were seen as not legitimate (and healthcare act was almost the only thing) (book by

Electricity you could regulate locally…

Britain will be loser with brexit, in relatively thoughtful process of Europe, GDPR and case against google.

China — empowered, legitimate state authority, demonstrates ways to constrain and limit free run of digital giants, both home grown and international.

Phil H — agree US is off the list of potential regulatory guides. One final act may be around DRM. Material devices and our rights — even if the EU develops strong privacy regulations, that the US still leads on design of the tech (even if made elsewhere). If US designed consumer goods come with DRM, like cars and coffee cups, our privacy will be US-defined not european.

David: Government/governance of the things, by the things, for the things. Which should we be most worrying of the last two?

Bill — driverless cars. The issue is of agency. When we kill someone as a driver, there’s objective tests which are almost bound to be applied. were you drunk, or having a heart attack? No question of agency in these things. With driverless cars, in aggregate, may kill a lot fewer people. But when the first one kills the first person, where does agency lie? The coder, the integrator, the owner? It will be litigated. Different answers in different regimes. This will become more widespread as more things become empowered!

Not about worrying about artificial general intelligence. Machine learning is constrained, fragile, brittle, as deployed today. Great successes where there’s exogenous rules, like Chess, Go, Warcraft. But where interaction is about negotiation — like in this panel — what the words mean, when we talk about Iot or digital agency — we are a long way from having that conversation with a machine, usefully.

David — Governance — By things for people, or by people for things? This may be a real choice.

Phil — governments, and governance. Governments don’t manage 100% population, and are bounded, mostly in cities. Not so much in slums, rural areas around the world. Trick may be to imagine a world with governance systems, created through our tech, to be fair and address inequalities, which we design thus, which apply where governments don’t.

There are opportunities around how we currently constitute government/governance in this future.


Q: sweeping discussion about need for new global governance to create shared values and standards. Is this useful?

A: Phil — believe in universal declaration of human rights. These are the universals. There are implementation questions. Universals are not represented by tech as designed. We need to stay on message that there are universal human rights. Chinese unlikely to build their IOT to encode those principles.

John — looks like there will be no opt out of the hypernetworked IOT world. Terribly tech determinism — rolling down hill — no opt out and its coming! Right to a private life seems unmaintainable in a world where almost everything we do is monitored (in such a way that can be monetised, actioned, etc). Several levels at which tech poses puzzles we can’t answer today. irresistible force, immovable object…. (will talk later about GDPR). But can a world like this be subject to any sort of democratic control? Or only by authoritarian states? Hence creation of attacking business model and implicit ideology. We are not capable of managing and controlling this tech.

Bill — we have a number of international governance structures, eg ITU (protocols), WTO (policy / globalisation). All are subject to capture, particularly where the state has been most fully delegitimised (USA at present). trade negotiations decoupled from interests of balance or fairness — eg transpacific — the core of this was a legitimisation of corp power, requiring disputes between investors and states to be settled via arbitration rather than litigation. How much of what had been civil litigation has been displaced to arbitration (eg employer vs employee) in US? (orthogonal to business models and unavoidability of digital tech, but informs it) This political discussion informs any regulatory responses. Finding a reason to bring the state back in will be critical issue. Won’t know whether there is a response until that begins. Europe — less ineffective though…

Q: The device paradigm; we selectively view tech, so we talk about the Things in the IOT, not the internet. A lot of the things we thought we owned are loyal to corporations not us — TV, car, alexa. We choose not to observe those relationships and instead fetishise the device. Miss the story!

Q: this is a very developed world conversation. How much of a leapfrog for less developed to invent new tech and infrastructure, closing gap?

Q: Foucault on relationship between knowledge and power. Heatmap of world connectivity — huge darknesses in africa, middle east. This is not a universal phenomenon. But also: these nodes and networks already transcend nations. Darkness in rural US, too. challenges are already beyond national governments. How can we use tech to empower those not empowered today.

A: Phil — don’t believe in leapfrog. poor countries invest in tech that others have developed, and profits accrue in the North. not a lot of evidence that the economic benefits are concrete enough to suggest leapfrogging. A moment as tech rolls out, where social inequality can be preserved — devices are a loss leader for data, and the data to drive design will come from large urban centres in the north. New devices will land and maybe redesign for developing world, repurposing outside wealthy cities. But demonstrable effects from IOT appearing first in rich cities and next wave design based on that.

Bill — centre of gravity and value in all things technological went from hardware to software, and now to data. Engine of monopolisation of value from data is an additional new, more than just economies of scale and network externalities etc that we are used to. Those things were big in days of telephony, AT&T customers called AT&T customers. That more data enables better algorithms which enables better service is a completely different fourth engine of monopolisation. In less developed countries, things are happening. eg Indonesia IT-enabled bikesharing — that’s a battle not just in china. (a Sargasso sea of abandoned bikes in Beijing) tech having a notable impact, even with things like bikes, without much infrastructure. Don’t believe in leapfrogging — eg MPesa in kenya was innovative but skewed by consumer power elsewhere and already over as a lead.

David: devices look like ours but they are someone else’s agents. There must be historic examples, but this time, the devices are dressed up as enabling, empowering things.

John — control is an integral part of fetishisation. iphones — do you think of it as yours? There’s nothing goes on your iphone that apple hasn’t approved. Books on Kindles can be remotely wiped (1984 example). Good example of rational humans losing their rationality — read internet on iphoneX — should these people be allowed to vote?? (joke) On developed world: look at Myanmar; recent arrival of cheap smartphones. They run facebook which gives them free data; and facebook has been used as a tool for ethnic cleansing — monetised in california. Kenya MPesa may be an execption, but if there’s profits to be made it’s usually in the global north. Narrative of tech being good for everyone is false.

Phil — people have used tech in creative ways, to provide specific governance goods. When finance doesn’t work in nigeria/kenya, no way to get small loan — now mobiles give ability to get that. Indonesia — when you see loggers coming, you can collect evidence of resources being taken on a phone, log forest destruction. Cause for optimism. these things in liminal spaces can advance governance.

Facebook: could contribute to democracy. POlitical scientists have argued that voting is a blunt instrument. interesting network experiments, say taking 200 representatives to meet experts and ideologues and deliberate for a week, come up with policy recommendations. civicly generated voter guides, processing evidence and boiling down issues to key things for public votes. Facebook could provide mechanisms like this, creative ways to execute democratic processes.

Bill — even before getting to that sophisticated level (eg in britain we’ve seen an even blunter instrument — single issue referendum!) — government digital service, taking simple lessons learned from consumer tech, apply to government/citizen interface. US copied it; code for america; great success in child benefit. Now applied to parole system, smartphone helps stop inadvertent breaches of conditions. Similar helpful things coming in medicine. Give hope.

John —Melvin Kranzberg — tech neither good nor bad nor neutral! The problem here is the business model. If you paid for facebook the dynamic would be totally different. With a different model we could avoid many problems and get closer to the utopia we’d like. IOT is following the current platform business models which is bad, need to find new models.

Bill — history of the problematic business model. Look at wireless — used to be point to point, until 1921, when wireless broadcast starts to appear. Hoover helped mobilise a patent pool (RCA) as a direct threat to the UK / Marconi patents which looked likely to dominate. RCA began to deliver tech to private sector, and search for business model. You can’t make people pay to receive a broadcast — SO hence advertising. Hoover denounced pollution of airwaves for public profit. Soap opera — origin in procter and gamble sponsorship. Popular novels were printed with ads, eg first edition Dickens. Supplement to reduce previous business model which was subscription to authors, to underwrite cost of printing. So — today’s business model not unique to tech today. It does have a unique history in non-rival goods — that’s where advertising has come from. Very hard to see how it gets disembedded. But we are seeing a rebirth of subscription models — Prime for instance. Evolution back to 1820s.

Q: I’m Vinay from the blockchain and i’m here to help. The problem is the data centre requiring capital, therefore economies of scale. So proposal: pay for use, open underlying networks governed by markets. Democratic processes to award money from capital pools to teams which offer modifications. OSS replaced Microsoft; a similar round of open source will replace facebook, once we figure out how to decentralise the underlying architecture. But can we protect the blockchain long enough to get there?

Q: issues with democratic choice, risk that citizens lose faith in democracy and own decision making capabilities… in similar way to how we give data to tech giants, will give decision making capabilities to algorithmocracy, as you can’t have decisions without info, and we have limited info (access, time, junk). cf Thinking Fast and Slow. Might we decide to trust machines, like the omniscient IOT network with AI?? Government of the people by the things :)

Q: will we be mobilised by Big Tech against our own governments?

David: what sets us free? what other historical links are there between politics, and things?

Q: things like climate change, an object of politics. can the IOT provide an infrastructure for understanding the objects of politics

John — optimistic side, that the IOT may surface things like climate change. Most rivers in the UK tweet their level, now… Is the business model sustainable? Perhaps not, long term…

Bill — remember how recent and fragile the coexistence of representative democracy and market capitalism is. And representative government we had before democracy. Burke — you elect me to exercise my judgement, and my judgement will be transparent (voting in parliament) to you, and if you don’t like it vote me out. Equivalent — need to see the training data for the algorithms; want to vote on who trains the algorithm that rules us all. Whatever the constituency is, believe in representative government — look at post-1945 italy and german constitutions, they don’t allow referenda, because simple/single issues are bad. (California is a state governed mostly by referenda.) Referenda are bad, almost as bad as government by algorithm.

David — democracy by design. But who votes out the designers? Can create spaces for deliberative democracy, but democracy isn’t just gathering public opinion, it’s also kicking the bastards out. That side might be withering away (or we get the hyperversion where we’re all each others bastards)

Phil — we should relax our assumptions that agency only lies with humans. Need nuanced ways we can talk about political action and the thing we study when we study social systems. Latour helpful here. Posit: we won’t get significant blockchains unless the state helps, public infra investment. (Dubai not big and important enough to count). State won’t disappear. Will be defined by its use of tech. A political actor more likely to choke on blockchain may be existing big tech. Would non net neutral networks (AT&T, netflix) support distributed ledgers? Would trust algorithm decisions if felt outcomes fair, maybe correctives for social inequalities… not afraid of governance goods by algorithm!

More after lunch here.