Category: datamining

Mozilla stops ad-network cookies

Mozilla, the developer of the Firefox web-browser, has decided that voluntary compliance by advertisers with its ‘Do Not Track’ settings is not working. Advertisers have basically been ignoring what is essentially a request by users, so instead of giving up, Mozilla has taken the right step and will simply not allow ad networks to install cookies on user’s computers or phones. This will of course cut ad revenue to some sites that rely on it, but it will also be a major step to slowing the proliferation of online tracking.

Of course, it can also be seen as a new negotiating position in a long conflict, as the Centre for Democracy and Technology points out, it could be a negotiating position that is all about trying to force companies to implement Do Not Track requests as a compromise from wholesale cookie-blocking. But I’m fully on board with Mozilla here either way. I very much doubt that Microsoft will take a similarly ethical stance on user control – because that’s what this is really about, not privacy as such but who has the right to control information about themselves.

The Unbearable Shallowness of Technology Articles… or, what Facebook Graph Search really means.

Wired has a feature article about Facebook’s new search tool. The big problem with it is that its vomit-inducing fawning over Facebook’s tech staff. In trying to make this some kind of human interest story – well, actually the piece starts off with Mark Zuckerberg’s dog, you see, he is human after all – of heroic tech folk battling with indomitable odds to create something amazing – what in science fiction criticism would be called an Edisonade – it almost completely muffles the impact of what a piece like this should be foregrounding, which is about what this system is, what is has been programmed to do and where it’s going.

And this is what Graph Search does, very simply: it is a search engine that will enable complex, natural language interrogation of data primarily but not limited to Facebook. So instead of trying to second-guess what Google might understand when you want to search for something, you would simply be able tell you what you ask. And because this is primarily ‘social’ – or about connection, and you should have already given up enough information to Facebook to enable it to ‘graph’ you so that it knows you, the results should supposedly be the kind if things you really wanted from your query. Supposedly. An FB developer in the article describes this as “a happiness-inducing experience” and further says, “We’re trying to facilitate good things.” However what this ‘happiness’ means, just like what ‘friendship’ means in the FB context, and what “good” means, just like the use of ‘evil’ in Google’s motto, is rather different than how we might understand such a term outside these contexts.

In the article, one example demonstrated by the developer is as follows:

[He] then tried a dating query — “single women who live near me.” A group of young women appeared onscreen, with snippets of personal information and a way to friend or message them. “You can then add whatever you want, let’s say those who like a certain type of music,” [he] said. The set of results were even age-appropriate for the person posing the query.

So when Mark Zuckerberg is quoted in the article saying that Graph Search is “taking Facebook back to its roots”, he seems to mean creeping on girls, as was, let us not forget, the main intention of the early Harvard version. Doesn’t this generate exactly the concern that the notorious ‘Girls Around Me’ app encountered? As the title of my favourite tumblr site has it, this isn’t happiness. Or it’s the happiness of the predator, the pervert and the psychopath.

But more fundamentally, this isn’t about privacy, or even online stalking. In fact, in many ways, both are side-issues here. This is about control and access: control over my information and how I access other information, not just on Facebook but in general. To me, the plans outlined for Graph Search look worrying, even outside of my idea of what would constitute happiness, because they have nothing to do with how I use Facebook or how I would want to use it. I don’t use Facebook as my gateway to the Web and I am never going to. As Eli Pariser pointed out in The Filter Bubble a couple of years back, that would both be limiting of my experience of the Web (and increasingly therefore of my communications more broadly) and give one organisation way too much power over both that experience and the future of the Web. But this does seem to be how Facebook wants it to be, and further, I suspect that, just like Bill Gates before him with his .NET initiative and other schemes, and just like the walled garden locked-in hardware that Apple produces, Zuckerberg is more interested in Facebook colonizing the entire online experience, or layering itself so entirely, tightly and intimately over the online world that the difference between that world and Facebook would seem all but invisible to the casual user.

These developments are dramatic enough in themselves. Never mind fluffy stories of heroic techies and their canine sidekicks.

Night of the Surveillance Dead

In one of those curious synchronicities that occasionally emerge out of the chaotic foam of the internet, I came across two stories (of an entirely different nature) featuring surveillance and ‘zombies’ this week.

The first is one that Ars Technica first publicized recently – the creation of new undeletable cookies. Cookies, for the still unaware, are little bits of code that sit on your computer and store information, usually relating to websites you have visited – so, passwords and the like. Originally they were simply a tool to make it easier to handle the proliferation of sites that needed login details from users. And in most cases, they used to be both moderately consensual (i.e. you would be, or could be, asked if you wanted to have you computer download one) and relatively easy to remove. However, in recent years, this has changed. For a start there are so many sites and applications using cookies that it has become inconvenient to ‘consent’ to them or to manage them in any unautomated way. The new development however is a system that uses the database capabilities in HTML5 rather than being a traditional cookie. The major problem with this, and you can read more about the technical details in the story, is that these cannot ever be deleted by the user, as when they are deleted, they respawn themselves, and recreate the data profile of the user by reaching into other areas of your computer (and even stuff you thought was also deleted). The company concerned, Ringleader Digital, which specializes in ‘targeted, trackable advertising’ for ‘real-time visibility’, says users can ‘opt-out’ by using a form on their website, but this so-called ‘opt-out’ is hedged about with terms and conditions.

Now, Ars Technica reports that an open-source developer, Samy Kamkar, has created ‘evercookie‘, a virtually indestructible cookie designed as an educational tool to make users aware of the presence of these new internet zombies that do their master’s bidding. It’s a neat idea but I wonder – and I hope you will excuse my taking the zombie metaphor just a little further here – whether in raising the dead to show that necromancy is bad, good wizards like Samy Kamkar might in the end just be contributing to the problem. It isn’t as if most ordinary users understand these strange powers. Perhaps the people who need to witness the power of these occult rites are the regulators. It’s not clear to me whether these kinds of programs would be considered in any way legal in most places with strong data-protection and privacy laws, like Canada and the EU – as the controversy over the similar British Telecom system, Phorm, showed. So I would be very interested in what the Canadian Privacy Commissioner has to say about it, for example. I will be asking them.

(The second zombie story I will add later…)

Surveillance, Coercion, Privacy and the Census

There’s been a huge furore here in Canada about the current government’s decision to abolish the long-form census. I’ve been following the debate more interested in what the proponents and opponents have been saying about privacy and surveillance rather than intervening. But it’s about time I got off the fence, so here’s my two cents’ worth. It may come out as an op-ed piece in one of the papers soon, I don’t know…

Sense about the Census:

Why the Long-form Census debate really matters.

The debate about the scrapping of the long-form census is in danger of being unhelpfully polarized. The result can only benefit the current government to the long-term detriment of the Canadian people. On the one hand, some of those campaigning for the reinstatement of the survey have dismissed issues of surveillance and privacy. On the other hand, supporters of its abolition have referred to ‘privacy’ and ‘coercion’ as if these words in themselves were reason enough to cut the survey. But the whole way in which privacy has been discussed is a red herring. We need to reaffirm a commitment to privacy alongside other collective social values not in opposition to them. We need privacy and we need the census.

First, coercion. The long-form census is undoubtedly a form of coercive state surveillance. One only has to glance at the recent history of state data collection and its role in discrimination and mass-murder to see that that one can be far too blasé about the possibility of states misusing statistics. Examples abound from the Holocaust to the genocide in Rwanda, and there is no reason to suppose that this could never happen again. In fact technology makes discrimination easier and more comprehensive: with sophisticated data-mining techniques, inferences can be made about individuals and groups from disparate and seemingly harmless personal data.

However, just because censuses have the potential for abuse, this does not make them wrong. Surveillance forms the basis of modern societies, good and bad, and coercion is all around us from the time we are children told by our parents not to play on the stairs. Coercion can be caring, protect us and improves our lives. The long-form census would have to be shown to be unfairly coercive, or not have enough beneficial policy outcomes to justify any coercion. This, the government has failed to do, whereas the campaign for the restoration of the survey has highlighted numerous examples of improvements in communities across Canada resulting from long-form census data.

Now to privacy. The campaign to restore the long-form census has seen frequent instances of the argument, ‘nothing to hide, nothing to fear’. This is one of the most glib arguments about privacy and surveillance, not only because of the potential abuse of state data collection but also because it assumes so much about what people should want to keep private. Another common argument is that privacy is irrelevant because ‘everyone gives away their personal information on Facebook anyway’. But the fact that some people chose to share parts of their lives with selected others does not imply that any infringement of privacy is acceptable. Privacy depends on context. Social networking or marketing trends do not mean that ‘anything goes’ with personal data.

In making these arguments, campaigners end up unwittingly bolstering a government strategy that relies not only on the evocation of ‘coercion’ but on pitting individual privacy against collective social goals. Yet, the government’s position is misleading. Privacy is not simply an individual right but also a collective social value. And further, just because the data is collected from individuals by the state, does not mean that the state infringes on privacy. It depends on whether the data is stored without consent in a way that identifies individuals or is used in a way negatively impacts upon them.

However, Statistics Canada have demonstrated a commitment to privacy within the census process. The long-form census data is not used to identify or target individuals. It is aggregated and used for wider community purposes. As Statistics Canada say quite on their website: “No data that could identify an individual, business or organization, are published without the knowledge or consent of the individual, business or organization.” The census returns are confidential and Statistics Canada employees are the only people who will ever have access to the raw returns, and they are bound by The Statistics Act. All this was confirmed by the Office of the Privacy Commissioner of Canada, who found the 2006 census fully compliant with privacy law.

So both privacy and coercion are red herrings. The conduct of the long-form census has demonstrated a commitment to privacy alongside other collective social values in support of individuals and the wider community. This moderate, sensible and profoundly Canadian position is now under threat. That is why this debate matters.

CIA buys into Web 2.0 monitoring firm

Wired online has a report that the US Central Intelligence Agency has bought a significant stake in a market research firm called Visible Technologies that specializes in monitoring new social media such as blogs, mirco-blogs, forums, customer feedback sites and social networking sites (although not closed sites like Facebook – or at least that’s what they claim).  This is interesting but it isn’t surprising – most of what intelligence agencies has always been sifting through the masses of openly available information out there – what is now called open-source intelligence – but the fact is that people are putting more of themselves out their than ever before, and material that you would never have expected to be of interest to either commercial or state organisations is now there to be mined for useful data.

(thanks, once again to Aaron Martin for this).

US Congress debates online data protection

The US House of Representatives will finally get to debate whether online advertising which tracks the browsing habits of users is a violation of privacy and needs to be controlled. A bill introduced by Rep. Rick Boucher of Virginia will be propsing an opt-out regime that gives users information about the uses to which their data will be put, and allows them to refuse to be enroled. At present many such services work entirely unannounced, placing cookies on users’ hard drives and using other tracking and datamining techniques, and without any way in which a user can say ‘no’. Of course, we have yet to see the results of the inveitable industry scare-stories and hard-lobbying on the what will be proposed, let alone pased. But the proposal itself is particularly significant because so far the US has so far always bowed to business interests on online privacy and data protection, and if this bill is pased, it is a sign that what EFF-founder, Howard Rhiengold, long ago called the ‘electronic frontier’ might start to acquire a little more law and order in favour of ordinary people.

Tokyo Brandscaping and the SuiPo system

Brandscaping is a term used in marketing to describe the metaphorical landscape of brands (either for a particular brand, company or sector), however it is also being used by some researchers, including me, to describe the way in which brands are being infiltrated into urban landscapes, with the ultimate aim of being ‘inhabitable’ perhaps even 24/7 (see for example Disney’s move into urban development with Celebration in Florida).

Contemporary brandscaping makes use of new ambient intelligence, pervasive or ubiquitous computing technologies (‘ubicomp’) and ubiquitous wireless communications to create a landscape in which the consumer is targeted with specific messages directing them to certain consumption patterns. Such communication cans of course be two-way and provide corporations with valuable and very personal data on consumption patterns. As I’ve argued in many presentations over the last few years, ubicomp is necessarily also ubiquitous surveillance (what I call ‘ubisurv’ – hence the name of this blog!) because to work it requires locatability and addressability. Japan, and Tokyo in particular, has been the site for a number of cutting edge experiments in this regard, including the ‘Tokyo Ubiquitous Technology Project’ which embedded 1000 RFID tags which can communicate with RFID-enabled keitai (mobile phones) in upscale Ginza as well as several other pilot schemes around Ueno Park and Shinjuku.

TUTP is not all about marketing surveillance however, part of the scheme has involved ‘Universal Design’ (UD) principles, with one experiment to embed chips in the yellow tactile tiles designed to help guide sight- and mobility-impaired people around the city so that useful access information could be passed through specially-enabled walking sticks. I’m very interested in such experiments as they indicate an alternative direction for ubicomp environments which are about genuinely enabling people who are currently disabled by social and architectural norms, and creating a richer sensory landscape. They show that both surveillance and ‘scary’ technology like RFID chips can be humanised.

Unfortunately in our consumer-capitalist world (and Tokyo is the exemplary city of hyper-consumption), marketing and building brandscapes tends to take priority over enabling the excluded and the disadvantaged. But there are different ways of doing this too, which can be more or less intrusive and consensual. The other day I was talking about the growth in functionality of the Suica smart travel card system. Suica-enabled keitai can now, be used buying all sorts of things and since 2006 there have been a growing number of ‘SuiPo’ (short for ‘Suica Poster’) sites, Suica-enabled advertising hoardings that will, on demand send information to your mobile e-mail address with on particular advertising in which you are interested if you pass your Suica card or phone over a scanner placed next to the poster (see photos below)

The difference between SuiPo and the Ginza RFID scheme however is that it with SuiPo is that it is the consumer who makes the choice whether to activate any particular poster’s additional information system. In this sense it is a development of the i-Mode system in which many keitai can read information from special barcodes embdedded in magazine advertisements. It doesn’t automatically call your phone every time you pass an enabled poster, once you have signed up. Not as high-tech but slightly more consensual. However this will, of course, lead to the accumulation of a lot of data on consumption interests. This potentially generates a massive consumer surveillance tool, because it can be linked up travel patterns (your registered Suica card sends information back on where you go – I was wrong about the absolute differences between London’s Oyster and Tokyo’s Suica systems the other day) and information about consumption.

So will this potential become reality? The page on privacy and data protection on the SuiPo website (as usual the link is hidden away at the bottom of the front page!), is pretty standard stuff except for the legitimate purposes for which the data can be used once you sign up. They are, for those who don’t read Japanese, for:

  1. Sending the specific requested information to you;
  2. Improving services;
  3. Data processing and analysis;
  4. JR East’s promotional marketing; and
  5. JR East customer questionnaires.

Purposes 2 and 3 pretty much allow JR to do anything it likes with the data once you have signed up, and there is no statement as to what can or cannot be done with data once it has been ‘mined’ – analysed and transformed into more useful to the company or other organisations (corporate or state) which might want to buy or access such knowledge. ‘Ubisurv’ indeed…

Google: ‘give us data or you could die!’

I’ve been keeping a bit of an eye on the way that online systems are being used to map disease spread, including by Google. What I didn’t anticipate is that Google would use this as a kind of emotional blackmail to persuade governments to allow them as much data as they like for as long as possible.

Arguing against the European Commission’s proposal that Google should have to delete personal data after 6 months, Larry Page claims that to do so would be “in direct conflict with being able to map pandemics” and that without this the “more likely we all are to die.”

Google talk a lot of sense sometimes –  I was very impressed with their Privacy counsel, Richard Fleischer, at a meeting I was at the other week – and in many ways they are now an intimate part of the daily lives of millions of people, but this kind of overwrought emotionalism does them no favours and belies their moto, ‘don’t be evil’.

(again, thanks to Seda Gurses for finding this)

FBI data warehouse revealed by EFF

Tenacious FoI and ‘institutional discovery’ work both in and out of the US courts by the Electronic Frontier Foundation has resulted in the FBI releasing lots of information about its enormous dataveillance program, based around the Investigative Data Warehouse (IDW). 

The clear and comprehensible report is available from EFF here, but the basic messages are that:

  •  the FBI now has a data warehouse with over a billion unique documents or seven times as many as are contained in the Library of Congress;
  • it is using content management and datamining software to connect, cross-reference and analyse data from over fifty previously separate datasets included in the warehouse. These include, by the way, both the entire US-VISIT database, the No-Fly list and other controversial post-9/11 systems.
  • The IDW will be used for both link and pattern analysis using technology connected to the Foreign Terrorist Tracking Task Force (FTTTF) prgram, in other words Knowledge Disovery in Databases (KDD) software, which will through connecting people, groups and places, will generate entirely ‘new’ data and project links forward in time as predictions.

EFF conclude that datamining is the future for the IDW. This is true, but I would also say that it was the past and is the present too. Datamining is not new for the US intelligence services, indeed many of the techniques we now call datamining were developed by the National Security Agency (NSA). There would be no point in the FBI just warehousing vast numbers of documents without techniques for analysing and connecting them. KDD may well be more recent for the FBI and this phildickian ‘pre-crime’ is most certainly the future in more ways than one…

There is a lot that interests me here (and indeed, I am currently trying to write a piece about the socio-techncial history of these massive intelligence data analysis systems), but one issue is whether this complex operation will ‘work’ or whether it will throw up so many random and worthless ‘connections’ (the ‘six-degrees of Kevin Bacon’ syndrome) that it will actually slow-down or damage actual investigations into real criminal activities. That all depends on the architecture of the system, and that is something we know little about, although there are a few hints in the EFF report…

(thanks to Rosamunde van Brakel for the link)

Phorm philling

UK satirical magazine, Private Eye, this week brings the ludicrous Stop Phoul Play website to my attention. This is a corporate spin site devoted entirely to defending BT’s underhand and intrusive ‘Phorm’ online advertising technology against what it calls ‘privacy pirates’ who they claim are either being paid or pushed to damage BT.

Those listed as ‘piracy pirates’ include the excellent investigative IT journal, The Register, the Open Rights Group and the brilliant Foundation for Information Policy Research (FIPR), along with numerous bloggers and contributors to web forums. Now, it may be that some other corporations with rival technologies would like Phorm to fail, just as Microsoft probably enjoys it a great deal every time Google takes a PR hit (or vice-versa), but to suggest that everyone who make a criticism of Phorm is secretly part of some conspiracy against BT is frankly, either stupid paranoid.

And there are very good reasons for being critical of Phorm in the trojan-like manner of its operation and the way in which it has been tested without the consent of users. As Private Eye also reminds us, Phorm has landed the UK government in legal trouble with the EU. It hardly needs a conspiracy to make people justifiably annoyed.

This is one of the weirder exercises in PR I have seen, not least because its paranoia and promotion of conspiracies can only be damaging to BT. Thus it is no surprise to find that, according to the The Register, that it is the product of the fevered imagination of Patrick Robertson, whose previous clients include the lovely General Pinochet and former Tory MP and convicted liar, Jonathan Aitkin. So go take a look at Stop Phoul Play (while it still exists…) – it really is quite insane.