UK/Japan Workshop on IoT

I recently represented PETRAS at a workshop organised to form links between UK and Japanese academics working on the Internet of Things held at the British Embassy in Tokyo 7-8th Sept 2017. There were approximately 12 academics from both countries who presented their work around the Internet of Things from a wide range of perspectives. One on my main observations is that the UK presented a much more socio-technical view compared with our Japanese counterparts which perhaps reflects the strength of interdisciplinary work in UK research. It was also interesting as to how much of the work in Japan is addressing ageing. With the fastest aging population in the world coupled with low birth rate this represents a huge issue going forward. The whole event was very interesting and extremely useful to have an insight on how different cultures are addressing the Internet of Things.

My slides are below followed by the transcript of my talk for those interested.

Slide 1.

This work is part of activities related to PETRAS Internet of Things Research Hub  which is a consortium of nine leading UK universities explore critical issues in privacy, ethics, trust, reliability, acceptability, and security related to the internet of things. These themes are considered across a range of IoT application areas such as wearables, home, city, health, transport, control etc but here I am considering the home as it is a useful way of engaging with a wide audience.

Slide 2.

If we are interested in Adoption of IoT the Hype Cycle provides a useful rhetorical tool for considering how this can be done. Whilst in some cases this curve suggests all technologies will eventually find a market this is not the case and some will never emerge from the chasm at the bottom of trough of disillusionment whilst others require multiple hypes (VR) before the may be adopted. In academic research Engineering and CS typicall operate on emerging technologies and whilst producing interesting prototype solutions rarely consider the societal changes/impact these technologies will require or produce.

 Slide 3.

The x-axis represents time from pasts to the futures (the use of the plural is deliberate to acknowledge these will be perceived differently by different individuals). The y-axis intersects the x-axis timeline at the present with a range from possible to impossible with actual centred on the present. Note that actual represents the point of commercial domestication of technology. Design Fictions and Vapourware/Vapourworlds seek to create a form of temporal shift of potential near and far future points of domestication so that they may be more readily understood in the present. Technologies that fail to reach the commercial domestication point could be considered as Lost Futures and their potential might be reconsidered by writing alternate histories.

Vapourware is a term commonly used to describe software and hardware that is announced, sometimes marketed, but is never actually produced. Vapourworlds is a term we use for projects are used by technology organizations to showcase future worlds in which these organizations and their products, are the enablers of preferable futures.

What contrasts Design Fiction from Vapourware and Vapourworlds is the intentionality underlying why these future visions are created. Vapourworlds are produced, predominantly by corporations, whose primary concern is in generating profit both now and in the future. While notions of profitability, or considerations related to profitability, may well form part of a Design Fiction world in order to make the fiction appear plausible, increasing a corporation’s ability to be profitable is rarely the intention behind the creation.

Coulton, P & Lindley, JG 2017, ‘Vapourworlds and Design Fiction: The Role of Intentionality”, no. Suppl. 1, pp. S4632-s4642. DOI: 10.1080/14606925.2017.1352960.

Slide 4.

In order to see this approach in action we shall use it to focus on privacy in IOT.

The Cayla Privacy policy… Says the voice data can be given to third parties, and to people they have a relationship with ‘only for specific purposes’ but, in fact that can be any purpose… Although it is ‘legal’ (apart from in Germany) it is nonetheless highly problematic because this particular bit of the IoT is for kids. There were also numerous security issues with this device.

Vizio.. Effectively shipped with spyware (in that they did not reveal this to owners) gathering usage data, which they sold to advertisers

Mirai targeted (ameras, routers, printers, some other hardware, to do DDoS atttacks. Relied on old hardware still with default passwords that had not been changed

Slide 5.

Whilst some people argue IoT needs more HCD to tackle this through behaviour change we argue HCD may be part of problem particularly drive to make things ‘simple for user.

Slide 6.

If we consider the example of a so0called smart thermostat what is its purpose?

From the user perspective it is simply a convenient way to control their heating.

From the thermostats perspective it purpose is to gather interaction data on heating which it stores in the cloud of its manufacture. However this is not clear in the interaction only in the T&Cs. HCD effectively produces invisibility shield in name of simplicity and as in previous examples could ultimately prove problematic.

Slide 7.

IoT Design needs to both take account different perspectives at play and make sure these are easily understood.

Slide 8.

Another point to raise is that connected devices are not the only things that influence IOT interactions.

Slide 9.

Business models, regulations, algorithms etc all play there part and effectively shape the design process.

We therefore argue that a constellation metaphor is the most useful for IOT. Describing IoT as constellations we suggest acknowledges that the appearance, utility, significance, and ‘meaning’ of network-nodes varies significantly depending on the observer’s perspective and each member of the constellation can have their own individual perspective.

Slide 10.

This is means we are effectively considering IOT from and OOO perspective whereby the human is simply another actant in the network and no more important than any other.

Lindley, JP & Coulton, P 2017, “Why the Internet of Things needs Object Orientated Ontology’ The Design Journal, vol 20, no. Suppl. 1, pp. S2486-s2857. DOI: 10.1080/14606925.2017.1352796.

Slide 11.

Now lets consider an example. How might this approach consider the oncoming GDPR regulation which WILL affect IOT design.

The main rights for individuals under the GDPR will be:

subject access,

to have inaccuracies corrected,

to have information erased,

to prevent direct marketing

to prevent automated decision-making and profiling, and data portability

User must make meaningful consent

This latter point we consider through the following Design Fiction

HT to Lachlan D. Urquhart who developed the cards shown in this image.

Slide 12.

Consider example of adding a smart lock to you not system as shown in video. Whilst we could have used a voice assistant to explain regulation as it was detected and installed it it too limiting in getting the notion of constellation effects across. Here we use a voice assistant in conjunction with an app that aims to reveal consequences of operation on security and that operating context can change and privacy should be able to be changed and renegotiated.

HT to Joe lindley and Hader Ali who did heavy lifting on this project.

Lindley, JG, Coulton, P, Akmal, H & Knowles, BH 2017, ‘Anticipating GDPR in Smart Homes Through Fictional Conversational Objects’.

Slide 13.

The app display differently levels of functionality against how much data is revealed about the user and how far it travels. The three circles represent levels of home, provider, and 3rd parties. The sharpness or blurriness represents how identifiable the user might be at these levels.

This is one simple example of how GDPR might be addressed and the purpose of the Design Fiction is to start conversations with designers and developers how they might approach it.

Slide 14.

Questions