Menu

Articles

Ethics

Could Counterfactuals Explain Algorithmic Decisions Without Opening the Black Box?

Previous article Next article
by David Sutcliffe

Exploring the role of algorithms in our everyday lives, and how a “right to explanation” for decisions might be achievable in practice

by David Sutcliffe

Algorithmic systems (such as those deciding mortgage applications, or sentencing decisions) can be very difficult to understand, for experts as well as the general public. Image: Ken Lane (CC BY-NC 2.0).

The EU General Data Protection Regulation (GDPR) has sparked much discussion about the “right to explanation” for the algorithm-supported decisions made about us in our everyday lives. While there’s an obvious need for transparency in the automated decisions that are increasingly being made in areas like policing, education, healthcare and recruitment, explaining how these complex algorithmic decision-making systems arrive at any particular decision is a technically challenging problem—to put it mildly. In their article “Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPR” which is forthcoming in the Harvard Journal of Law & Technology, Sandra Wachter, Brent Mittelstadt, and Chris Russell present the concept of “unconditional counterfactual explanations” as a novel type of explanation of automated decisions that could address many of these challenges. Counterfactual explanations describe the minimum conditions that would have led to an alternative decision (e.g. a bank loan being approved), without the need to describe the full logic of the algorithm. Relying on counterfactual explanations as a means to help us act rather than merely to understand could help us gauge the scope and impact of automated decisions in our lives. They might also help bridge the gap between the interests of data subjects and data controllers, which might otherwise be a barrier to a legally binding right to explanation. We caught up with the authors to explore the role of algorithms in our everyday lives, and how a “right to explanation” for decisions might be achievable in practice: Ed: There’s a lot of discussion about algorithmic “black boxes” — where decisions are made about us, using data and algorithms about which we (and perhaps the operator) have no direct understanding. How prevalent are these systems? Sandra: Basically, every decision that can be made by a human can now be made by an algorithm, which can be a good thing. Algorithms (when we talk about artificial intelligence) are very good at spotting patterns and…

Algorithmic systems (such as those deciding mortgage applications, or sentencing decisions) can be very difficult to understand, for experts as well as the general public. Image: Ken Lane (CC BY-NC 2.0).

Economics

A distributed resilience among darknet markets?

Previous article Next article
by Martin Dittus

The actions by law enforcement were deliberately structured to seed distrust in illicit trading platforms. Did this effort succeed?

by Martin Dittus

You may have seen the news earlier this year that two large darknet marketplaces, Alphabay and Hansa, have been taken down by international law enforcement. Particularly interesting about these takedowns is that they were deliberately structured to seed distrust among market participants: after Alphabay closed many traders migrated to Hansa, not aware that it had already covertly been taken over by the police. As trading continued on this smaller platform, the Dutch police and their peers kept track of account logins, private messages, and incoming orders. Two weeks later they also closed Hansa, and revealed their successful data collection efforts to the public. Many arrests followed. The message to illicit traders: you can try your best to stay anonymous, but eventually we will catch you. By coincidence, our small research team of Joss Wright, Mark Graham, and I had set out earlier in the year to investigate the economic geography of darknet markets. We had started our data collection a few weeks earlier, and the events took us by surprise: it doesn’t happen every day that a primary information source gets shut down by the police. While we had anticipated that some markets would close during our investigations, it all happened rather quickly. On the other hand, this also gave us a rare opportunity to observe what happens after such a takedown. The actions by law enforcement were deliberately structured to seed distrust in illicit trading platforms. Did this effort succeed? Let’s have a look at the data. The chart above shows weekly trading volumes on darknet markets for the period from May to July 2017. The black line shows the overall trading volume across all markets we observed at the time. Initially, Alphabay (in blue) represented a significant share of this overall trade, while Hansa (in yellow) was comparably small. When Alphabay was closed in week 27, overall sales dropped: many traders lost their primary market. The following week,…

Economics, Governance & Security, Mapping, Social Data Science

Mapping Fentanyl Trades on the Darknet

Previous article Next article
by Mark Graham

The US accounts for almost 40% of global darknet trade, with Canada and Australia at 15% and 12%, respectively.

by Mark Graham

My colleagues Joss Wright, Martin Dittus and I have been scraping the world’s largest darknet marketplaces over the last few months, as part of our darknet mapping project. The data we collected allow us to explore a wide range of trading activities, including the trade in the synthetic opioid Fentanyl, one of the drugs blamed for the rapid rise in overdose deaths and widespread opioid addiction in the US. The map shows the global distribution of the Fentanyl trade on the darknet. The US accounts for almost 40% of global darknet trade, with Canada and Australia at 15% and 12%, respectively. The UK and Germany are the largest sellers in Europe with 9% and 5% of sales. While China is often mentioned as an important source of the drug, it accounts for only 4% of darknet sales. However, this does not necessarily mean that China is not the ultimate site of production. Many of the sellers in places like the US, Canada, and Western Europe are likely intermediaries rather than producers themselves. In the next few months, we’ll be sharing more visualisations of the economic geographies of products on the darknet. In the meantime you can find out more about our work by Exploring the Darknet in Five Easy Questions. Follow the project here: https://www.oii.ox.ac.uk/research/projects/economic-geog-darknet/ Twitter: @OiiDarknet

Wellbeing

Why we shouldn’t be pathologising online gaming before the evidence is in

Previous article Next article
by David Sutcliffe

New study suggests that Internet Gaming Disorder (IGD) may not, in itself, be robustly associated with important clinical outcomes.

by David Sutcliffe

There are active debates surrounding Internet Gaming Disorder (IGD), however, a new study suggests that it may not, in itself, be robustly associated with important clinical outcomes.

Internet-based video games are a ubiquitous form of recreation pursued by the majority of adults and young people. With sales eclipsing box office receipts, games are now an integral part of modern leisure. However, the American Psychiatric Association (APA) recently identified Internet Gaming Disorder (IGD) as a potential psychiatric condition and has called for research to investigate the potential disorder’s validity and its impacts on health and behaviour. Research responding to this call for a better understanding of IGD is still at a formative stage, and there are active debates surrounding it. There is a growing literature that suggests there is a basis to expect that excessive or problematic gaming may be related to lower health, though findings in this area are mixed. Some argue for a theoretical framing akin to a substance abuse disorder (i.e. where gaming is considered to be inherently addictive), while others frame Internet-based gaming as a self-regulatory challenge for individuals. In their article “A prospective study of the motivational and health dynamics of Internet Gaming Disorder”, Netta Weinstein, the OII’s Andrew Przybylski, and Kou Murayama address this gap in the literature by linking self-regulation and Internet Gaming Disorder research. Drawing on a representative sample of 5,777 American adults they examine how problematic gaming emerges from a state of individual “dysregulation” and how it predicts health—finding no evidence directly linking IGD to health over time. This negative finding indicates that IGD may not, in itself, be robustly associated with important clinical outcomes. As such, it may be premature to invest in management of IGD using the same kinds of approaches taken in response to substance-based addiction disorders. Further, the findings suggests that more high-quality evidence regarding clinical and behavioural effects is needed before concluding that IGD is a legitimate candidate for inclusion in future revisions of the Diagnostic and Statistical Manual of Mental Disorders. We caught up with Andy to explore the implications of the study:…

There are active debates surrounding Internet Gaming Disorder (IGD), however, a new study suggests that it may not, in itself, be robustly associated with important clinical outcomes.

Ethics, Governance & Security, Interviews

Censorship or rumour management? How Weibo constructs “truth” around crisis events

Previous article Next article
by David Sutcliffe

Examining the content moderation strategies of Sina Weibo, China’s largest microblogging platform, in regulating discussion of rumours following the 2015 Tianjin blasts.

by David Sutcliffe

On 12 August 2015, a series of explosions killed 173 people and injured hundreds at a container storage station at the Port of Tianjin. Tianjin Port by Matthias Catón (Flickr CC BY-NC-ND 2.0).

As social media become increasingly important as a source of news and information for citizens, there is a growing concern over the impacts of social media platforms on information quality—as evidenced by the furore over the impact of “fake news”. Driven in part by the apparently substantial impact of social media on the outcomes of Brexit and the US Presidential election, various attempts have been made to hold social media platforms to account for presiding over misinformation, with recent efforts to improve fact-checking. There is a large and growing body of research examining rumour management on social media platforms. However, most of these studies treat it as a technical matter, and little attention has been paid to the social and political aspects of rumour. In their Policy & Internet article “How Social Media Construct ‘Truth’ Around Crisis Events: Weibo’s Rumor Management Strategies after the 2015 Tianjin Blasts”, Jing Zeng, Chung-hong Chan and King-wa Fu examine the content moderation strategies of Sina Weibo, China’s largest microblogging platform, in regulating discussion of rumours following the 2015 Tianjin blasts. Studying rumour communication in relation to the manipulation of social media platforms is particularly important in the context of China. In China, Internet companies are licensed by the state, and their businesses must therefore be compliant with Chinese law and collaborate with the government in monitoring and censoring politically sensitive topics. Given most Chinese citizens rely heavily on Chinese social media services as alternative information sources or as grassroots “truth”, the anti-rumour policies have raised widespread concern over the implications for China’s online sphere. As there is virtually no transparency in rumour management on Chinese social media, it is an important task for researchers to investigate how Internet platforms engage with rumour content and any associated impact on public discussion. We caught up with the authors to discuss their findings: Ed.: “Fake news” is currently a very hot issue, with Twitter and Facebook both…

On 12 August 2015, a series of explosions killed 173 people and injured hundreds at a container storage station at the Port of Tianjin. Tianjin Port by Matthias Catón (Flickr CC BY-NC-ND 2.0).

Democracy, Interviews, Politics & Government

Does Internet voting offer a solution to declining electoral turnout?

Previous article Next article
by David Sutcliffe

While a technological fix probably won’t address the underlying reasons for low turnout, it could help stop further decline by making voting easier.

by David Sutcliffe

Electronic voting in Brussels. © European Union 2014 – European Parliament.

e-Voting had been discussed as one possible remedy for the continuing decline in turnout in Western democracies. In their Policy & Internet article “Could Internet Voting Halt Declining Electoral Turnout? New Evidence that e-Voting is Habit-forming”, Mihkel Solvak and Kristjan Vassil examine the degree to which e-voting is more habit forming than paper voting. Their findings indicate that while e-voting doesn’t seem to raise turnout, it might at least arrest its continuing decline in Western democracies. And any technology capable of stabilising turnout is worth exploring. Using cross-sectional survey data from five e-enabled elections in Estonia—a country with a decade’s experience of nationwide remote Internet voting—the authors show e-voting to be strongly persistent among voters, with clear evidence of habit formation. While a technological fix probably won’t address the underlying reasons for low turnout, it could help stop further decline by making voting easier for those who are more likely to turn out. Arresting turnout decline by keeping those who participate participating might be one realistic goal that e-voting is able to achieve. We caught up with the authors to discuss their findings: Ed.: There seems to be a general trend of declining electoral turnouts worldwide. Is there any form of consensus (based on actual data) on why voting rates are falling? Mihkel / Kristjan: A consensus in terms of a single major source of turnout decline that the data points to worldwide is clearly lacking. There is however more of an agreement as to why certain regions are experiencing a comparatively steeper decline. Disenchantment with democracy and an overall disappointment in politics is the number one reason usually listed when discussing lower and declining turnout levels in new democracies. While the same issues are nowadays also listed for older established democracies, there is no hard comparative evidence for it. We do know that the level of interest in and engagement with politics has declined across the board in Western…

Electronic voting in Brussels. © European Union 2014 – European Parliament.

Wellbeing

From private profit to public liabilities: how platform capitalism’s business model works for children

Previous article Next article
by Huw Davies

Why has platform capitalism come to dominate children’s relationship to the internet and why is this problematic?

by Huw Davies

Young people choose to use platforms for play, socialising and expressing their identity. Image by Brad Flickinger (Flickr: CC BY 2.0)

Two concepts have recently emerged that invite us to rethink the relationship between children and digital technology: the “datafied child” (Lupton & Williamson, 2017) and children’s digital rights (Livingstone & Third, 2017). The concept of the datafied child highlights the amount of data that is being harvested about children during their daily lives, and the children’s rights agenda includes a response to ethical and legal challenges the datafied child presents. Children have never been afforded the full sovereignty of adulthood (Cunningham, 2009) but both these concepts suggest children have become the points of application for new forms of power that have emerged from the digitisation of society. The most dominant form of this power is called “platform capitalism” (Srnicek, 2016). As a result of platform capitalism’s success, there has never been a stronger association between data, young people’s private lives, their relationships with friends and family, their life at school, and the broader political economy. In this post I will define platform capitalism, outline why it has come to dominate children’s relationship to the internet and suggest two reasons in particular why this is problematic. Children predominantly experience the Internet through platforms ‘At the most general level, platforms are digital infrastructures that enable two or more groups to interact. They therefore position themselves as intermediaries that bring together different users: customers, advertisers, service providers, producers, suppliers, and even physical objects’ (Srnicek 2016, p43). Examples of platforms capitalism include the technology superpowers – Google, Apple, Facebook, and Amazon. There are, however, many relevant instances of platforms that children and young people use. This includes platforms for socialising, platforms for audio-visual content, platforms that communicate with smart devices and toys, and platforms for games and sports franchises and platforms that provide services (including within in the public sector) that children or their parents use. Young people choose to use platforms for play, socialising and expressing their identity. Adults have also introduced platforms…

Young people choose to use platforms for play, socialising and expressing their identity. Image by Brad Flickinger (Flickr: CC BY 2.0)

Economics, Interviews

Introducing Martin Dittus, Data Scientist and Darknet Researcher

Previous article Next article
by David Sutcliffe

Martin Dittus is a Data Scientist at the Oxford Internet Institute. The stringent ethics process governing his research means he currently can’t even contact anyone on the marketplace.

by David Sutcliffe

We’re sitting upstairs, hunched over a computer, and Martin is showing me the darknet. I guess I have as good an idea as most people what the darknet is, i.e. not much. We’re looking at the page of someone claiming to be in the UK who’s selling “locally produced” cannabis, and Martin is wondering if there’s any way of telling if it’s blood cannabis. How would you go about determining this? Much of what is sold on these markets is illegal, and can lead to prosecution, as with any market for illegal products. But we’re not buying anything, just looking. The stringent ethics process governing his research means he currently can’t even contact anyone on the marketplace. [Read more: Exploring the Darknet in Five Easy Questions] Martin Dittus is a Data Scientist at the Oxford Internet Institute, and I’ve come to his office to find out about the OII’s investigation (undertaken with Mark Graham and Joss Wright) of the economic geographies of illegal economic activities in anonymous Internet marketplaces, or more simply: “mapping the darknet.” Basically: what’s being sold, by whom, from where, to where, and what’s the overall value? Between 2011 and 2013, the Silk Road marketplace attracted hundreds of millions of dollars worth of bitcoin-based transactions before being closed down by the FBI, but relatively little is known about the geography of this global trade. The darknet throws up lots of interesting research topics: around traffic in illegal wildlife products, the effect of healthcare policies on demand for illegal prescription drugs, whether law enforcement has (or can have) much of an impact, questions around the geographies of trade (e.g. sites of production and consumption), and the economics of these marketplaces—as well as the ethics of researching all this. OII researchers tend to come from very different disciplinary backgrounds, and I’m always curious about what brings people here. A computer scientist by training, Martin first worked as a software developer…

Economics

Exploring the Darknet in Five Easy Questions

Previous article Next article
by David Sutcliffe

We caught up with Martin Dittus to find out some basics about darknet markets, and why they’re interesting to study.

by David Sutcliffe

Darknet marketplaces are typically set up to engage in the trading of illicit products and services, and are considered criminal in most jurisdictions. Image: Dennis Yip (Flickr).

Many people are probably aware of something called “the darknet” (also sometimes called the “dark web”) or might have a vague notion of what it might be. However, many probably don’t know much about the global flows of drugs, weapons, and other illicit items traded on darknet marketplaces like AlphaBay and Hansa, the two large marketplaces that were recently shut down by the FBI, DEA and Dutch National Police. We caught up with Martin Dittus, a data scientist working with Mark Graham and Joss Wright on the OII’s darknet mapping project, to find out some basics about darknet markets, and why they’re interesting to study. Firstly: what actually is the darknet? Martin: The darknet is simply a part of the Internet you access using anonymising technology, so you can visit websites without being easily observed. This allows you to provide (or access) services online that can’t be tracked easily by your ISP or law enforcement. There are actually many ways in which you can visit the darknet, and it’s not technically hard. The most popular anonymising technology is probably Tor. The Tor browser functions just like Chrome, Internet Explorer or Firefox: it’s a piece of software you install on your machine to then open websites. It might be a bit of a challenge to know which websites you can then visit (you won’t find them on Google), but there are darknet search engines, and community platforms that talk about it. The term ‘darknet’ is perhaps a little bit misleading, in that a lot of these activities are not as hidden as you might think: it’s inconvenient to access, and it’s anonymising, but it’s not completely hidden from the public eye. Once you’re using Tor, you can see any information displayed on darknet websites, just like you would on the regular internet. It is also important to state that this anonymisation technology is entirely legal. I would personally even argue that such…

Darknet marketplaces are typically set up to engage in the trading of illicit products and services, and are considered criminal in most jurisdictions. Image: Dennis Yip (Flickr).

Economics, Governance & Security, Interviews

Digital platforms are governing systems—so it’s time we examined them in more detail

Previous article Next article
by David Sutcliffe

It’s important that we take a multi-perspective view of the role of digital platforms in contemporary society.

by David Sutcliffe

Digital platforms strongly determine the structure of local interactions with users; essentially representing a totalitarian form of control. Image: Bruno Cordioli (Flickr CC BY 2.0).

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 formalised rules. In his Policy & Internet article “Platform Logic: An Interdisciplinary Approach to the Platform-based Economy”, Jonas Andersson Schwarz argues that digital platforms enact a twofold logic of micro-level technocentric control and macro-level geopolitical domination, while supporting a range of generative outcomes between the two levels. Technology isn’t ‘neutral’, and what designers want may clash with what users want: so it’s important that we take a multi-perspective view of the role of digital platforms in contemporary society. For example, if we only consider the technical, we’ll notice modularity, compatibility, compliance, flexibility, mutual subsistence, and cross-subsidisation. By contrast, if we consider ownership and organisational control, we’ll observe issues of consolidation, privatisation, enclosure, financialisation and protectionism. When focusing on local interactions (e.g. with users), the digital nature of platforms is seen to strongly determine structure; essentially representing an absolute or totalitarian form of control. When we focus on geopolitical power arrangements in the “platform society”, patterns can be observed that are worryingly suggestive of market dominance, colonisation, and consolidation. Concerns have been expressed that these (overwhelmingly US-biased) platform giants are not only enacting hegemony, but are on a road to “usurpation through tech—a worry that these companies could grow so large and become so deeply entrenched in world economies that they could effectively make their own laws.” We caught up with Jonas to discuss his findings: Ed.: You say that there are lots of different ways of considering “platforms”: what (briefly) are some of these different approaches, and why should they be linked up…

Digital platforms strongly determine the structure of local interactions with users; essentially representing a totalitarian form of control. Image: Bruno Cordioli (Flickr CC BY 2.0).

Close Menu