OTOCRON

This is basically the content of a talk I did for the Club of Amsterdam session on Ambient Intelligence, at February 22, 2007. I was asked to speculatively reflect on the future of ambient intelligence and the media. This is what I came up with, slightly fleshed out and clarified.

What impact will the development of Ambient Intelligence (AmI) have on the media ecology? In the AmI vision of the near future, many more things will start behaving like media - ‘everything will talk to everything’… What does that mean? And how will that affect the media ecology? I’ll break it down to a series of related questions and remarks on where we might find answers.

The wave of user generated content that’s currently happening, points to an increased democratisation of the media production process. (By the way: some people obviously still make better stuff than others - there is no equally impressive democratisation of talent going on.) The media ecology is also becoming increasingly interactive, increasingly 2-directional. Next to the internet we have on-demand content over TV cable networks, IPTV, digital radio, interactive television with settop boxes, the TIVO.

The powernodes in the media network have not ceased to exist, and are in fact strong as ever: BBC, TimeWarner, Reuters, Universal, CNN, etc. New powernodes have emerged: Nokia , YouTube, MySpace - yet others are emerging: ngo’s that become content providers, car navigation systems are becoming an important media channel. Next to this we have an immense and fastly growing amount of smaller media nodes - that mainly use internet. I’ll look at the relevance of this in the context of AmI later.

But when we talk about AmI, we talk not of The Media as we currently know them, but of new organisms in the media ecology. We talk of the way computronium , to quote Charles Stross among others, gets embedded in our physical surroundings. Many of those new things that become media are not media for humans in the first place - they are primarily media for other things - they are machine oriented media. They distill and traffic information: differences that make differences: sensor tresholds are surpassed, resulting in bitflips or datastreams, that trigger events in other locations and actuators. AmI means that many more things become media - and also that many more things will get mediated. I propose a new metric to describe this development consisely: the REVOS. (short for Reported Events per VOlume Second).

Through the ever increasing density of the media ecology and through the increasing amounts of user-generated content and through the distribution of all the new organisms in the media ecology and their machine-generated content the REVOS rate of the world will continue to increase rapidly. At some places more so than at others, obviously.

It is important to note that in the current proposals on AmI-like implementations, most of the distributed sensors. i/o boards and other smart matter only talks with some dedicated other smart stuff and that they don’t talk with the rest. So, some things will talk to some other things, rather then everything to everything… but maybe we can still do something about that. If we go along with the ideals of AmI: “environments that react intelligently to users needs” , the question arises how this array of different applications and functions will integrate to an environment that we can relate to. I’m thinking here of an urban environment, not of the space of one household. I’m thinking of an environment everyone can use. An environment that also leaves us also with an important sense of agency (otherwise, how can we trust it?)

Smart Matter & Ambient Intelligence
How to turn this smart matter into coherent:
- locative media?
- context sensitive media?
- taste sensitive media?
- fashion sensitive media?

One way to start unpacking that question could be to ask: how to integrate human generated content & non-human generated content on a general level? (We already have quite a bit of non-human generated content - streams from automated cctv systems with movement sensors, traffic sensors, RFID enabled logging systems, aircraft logging systems, weather data from measuring stations etc.)
Possible answers, that are also in line with the current changes in the media ecology could be the following: The first thing is that all these new media organisms should become part of the internet. They should have IP adresses and they should talk via the http protocol. This means that no major new infrastructures have to be developed. Furthermore, they should not be part of closed services, or applications. Like it has become normal procedure in the Web2.0 context, builders of new applications should continue to publish their API’s, allowing other parties to talk with their applications on the programming level.
But if we really go idealistic… These new media organisms will generate bits, numbers, characterstrings (text?), photo’s, sounds and video, but they should also generate nifty metadata, that allows aggregation of their data in various other human-desired contexts, the same metadata they use to organize their inputs from other sources. Standardisation of their metadata in various categories (relating to time, place, to the kind of mediaobject they are, or to the kind of aggregation they could belong to) would easily allow various devices and services to talk to each other (without too complex programming), and would easily allow new parties to develop new services using existing networks and data.
We would also need standard protocols for developing future kinds of metadata.

Tagcloud
From the human perspective, a proven useful way to organise some kinds of metadata to mediate between non-human centered media and the way humans understand is the tagcloud. The tagcloud is as semantically rich as human language, and at the same time sufficiently unambiguous for computers. I propose using tagclouds partly consisting of standardised metadata, and partly consisting of personal tags and choices as a simple way to organize human centered mediation of human and non-human generated data. Flock is an example of a browser that has already embedded ways to deal with different metadata coming from various Web2.0 services, to allow really simple aggregation by its users. Yahoo Pipes is another example of a generic aggregation tool for various web outputs.
Similar ideas would start to apply to our physical surroundings in my public space AmI scenario.

One-and-a-halfLife
AmI (and its current predecessors like locative media projects, GPS readers and RFID systems) fuses the physical surrounding with the digitial. This means that media content will extend further towards physical experiences, which implies that here and now becomes important at the expense of the new media paradigm ‘anytime, anywhere’.RFID, motes hives networks, IPv6, GPS awareness - all these techno-practices are already quickly establishing themselves as part of the media ecology. Also there are new media surfaces emerging: clothes, public screens, electronic paper. What will happen is that we’ll start to navigate physical space more like we navigate online information. We’ll move from a FirstLife next to SecondLife towards a permanent One-and-a-HalfLife (at least in urban environments)
The media ecology that includes this kind of general AmI will be more layered, with different layers inscribing different levels of reality. Myriads of new media niches emerge: from extremely local, personal, time- and context sensitive mediaflows, to worldwide time indepependent unpersonal mediaflows, and every possible combination thereof.All this will be organised through a combination of open API’s, standardised metadata and personal tagclouds, that behave like experience filters, or like smart advising librarians or more like hormones: using your testosteron tagcloud really gives a very different feel to the city than using your lavender tagcloud - for instance.
Users shape their personal paths through all available mediaflows, but these paths will partly overlap: powernodes in the medianetwork will still exist on global and regional level: olympic games, harry potter, bbc, the evening news, a soap in your own language, highbandwidth squares, public benches that are access points to funny personal experiences. Because people need to acknowledge the shape of reality to each other, they need something to orient themselves with, that allows them to share aspects of perceived reality with others.

So, to structure and facilitate the collective flow of media between non-humans among each other, between humans and non-humans and between humans through non-humans, we’d need open API’s and standardised types of metadata. Some are in place: GPS metadata and timebased metadata. Metadata models that relate to output channels (tv, mobile, billboard, loudspeaker, car navigator etc.) are being developed simultaneously by different brooadcast organisations and press organisations . There is no open standard… to my knowledge. Here is work to do. On top of the arrangements with controlled vocabularies of metadata, there should be room and functionality that supports user generated metadata, relating to times, places, events, objects, people animals, experiences etcetera. Sophisticated tagclouds that can structure tastes, fashions, moods, subcultures, and that can be shared at will by those interested to do so.

The Long Tail in meat space
Will we, humans, manage in this hybrid physical/digital ocean of (meta)dataflows? If so, how? Let’s take a contemporary evolutionary approach. We don’t talk about survival of the fittest and death to the rest, but: what can exist will exist - I mean: think of Media and the Long Tail.The Long Tail is here understood as the pattern by which the bulk of traffic on a network consists of an enormous amount of small flows between minor nodes instead of the big flows between big nodes.
IF media nodes and flows are online uniquely identifiable products, services and items.
AND IF the network is sufficiently big, connected and distributed
AND IF users (AGENTS!) are free to browse, engage, choose and discard these products services and items
THEN the Long Tail starts to apply.

So in this emerging AmI media ecology, the long tail will start to apply to all kinds of phenomena that also exist in physical reality. Think of the Long Tail in tourism: instead of being stuck with the Anne Frank House, the Heineken Brewery and the Albert Cuyp Market for a one day visit to Amsterdam, you could follow your highly peculiar tagtrail (namedtestoreron, or spinoza, or mother of pearl or whatever) through Amsterdam - a tagtrail built from experiences and metadata added by a relatively small group of totally unrelated world citizens that you’ve subscribed to. I’m talking about the del.icio.us or LastFM idea, applied to physical space, usable as you’re walking through town.

Issues and OpportunitiesTo allow people to find relevant behaviours in relation to these technologies, to make this a space for innovation, to allow this blended physical digital culture to bloom, there are a couple of issues to adress.
1 Open API’s and metadata standards (a policy issue, a design issue as well as philosophical issue) In my AmI scenario we’ll need classes of reliably translatable metadata in open libraries, open RFID systems, open ONS providers.This does have important philosophical implications: these classes of metadata will be powerful categories that henceforth organise our relation with the mediating/mediated environment (see also: transparancy) Interestign would be to see how people that prefer to use different classes realte to each other.
2 Trust. How can people trust such an environment? An environment that knows everything about them? Trust is manufactured through 1 transparancy and a feeling of 2 agency / negotiability on the part of users.
2.1 Transparancy (a policy issue and a design issue) Reported events are events of which the record should be publicly available, not only for some other things, but also for interested humans. In this scenario we’re outsourcing more and more parts of decision making processes to configurations of little tools and other people. We will want the ability to know and see how they work, we want to be able to see where what is happening. (even if we don’t often use this ability) Media transparancy is of course the counterpart of media literacy Basic network knowledge and protocols should be part of the curriculum of primary schools.
2.2 Agency( a policy issue and a design issue) Agency, the ability to cause and shape events, becomes more and more distributed over networks of humans and non-humans in an AmI scenario. The Long Tail pattern and the tagcloud are models that are potentially beneficial for a very large part of users. They imply important agency on the level of human users. The privacy issue I see as one embodiment of the agency issue. Privacy ceases to be an issue when it is negotiable. That implies agency.
3 And then there are the business oppportunities: Physical space browsers need to be developed for on mobile phones. There is demand for content aggregators of various nature: mashup providers, retailers of neatly packaged metadata aggregators and user-generated tagclouds (you land at Schiphol, your phone picks up the network, you get an offer for a kinky amsterdam tagtrail collected by a group of young japanese tourists, - for only euro 2,99 and then you get the Golden Age trail for free!) And of course there is room for a whole load of consultants.

As a closing remark: the above is partly fictional but not intended as exotic. And it is more value fiction than science fiction. We won’t need to be inventing new languages, but only new sentences and stories. Http will still work. We are already moving to a new IP regime (ipv6) that supports approximately 5 x 10 to-the-28th addresses for each of the roughly 6.5 billion people alive today, Object Name Services are just an add-on to a Domain Name Servers, and Physical Mark-up Language is just a kind of XML.