You are not logged in. LOG IN NOW >

How Technology Is and Isn't Helping Fight Corruption in India

BY David Eaves | Tuesday, May 28 2013

Sunil Abraham (photo: David Sasaki/Flickr)

I recently sat down with Sunil Abraham, the founder and executive director of the Center for Internet & Society (CIS) in Bangalore to talk about the center, and his views on the role of technology and openness in politics and society.

Abraham, who founded the CIS in 2008, has been active in the open source and technology space for almost two decades, constantly balancing research and theory, with a strong desire to get his hands dirty and actually make things happen. Prior to starting CIS he held both a Sarai FLOSS fellowship and Ashoka fellowship in which he explored the democratic potential of the Internet. However, before this he worked as a social entrepreneur and free software advocate, founding Mahiti, a company that implemented free software solutions for volunteer organizations. The company continues to thrive today employing over 50 engineers.

Today the Center for Internet and Society serves as a non-profit think tank and advocacy organization that research and explores policy options on freedom of expression, privacy, accessibility for persons with disabilities, access to knowledge and IPR reform, and openness. Over the past five years it has seen its influence grow as it becomes increasingly recognized for its expertise and critical view, within both government and the media, in India.

For part of the conversation I asked Abraham his thoughts on I Paid a Bribe, a website launched in August 2010 that allows users to report when and where they were asked to pay a bribe to a public official.

What follows is a fascinating critique of not just I Paid a Bribe but the UID (biometric ID program being rolled out across India):

Sunil Abraham:

The first complication with I Paid A Bribe is it is a quantitative approach to the problem.

Second, it assumes that those reporting the bribes are honest and don't have any other agendas.

Third, it also depends on the novelty effect. People can get bored when there is no feedback. The loop is not closed. To close the loop, some of the things that go on with anonymous reporting cannot happen, and to close the loop it almost needs to become a paralegal infrastructure. It has to talk to law enforcement and people have to be arrested, prosecuted and put away. To really address a problem like that is complex.

So it fails at each of those challenges.

The real challenge in India is high-ticket bribes that happen at the top of the pyramid. And those bribes are done in a very sophisticated fashion. Ordinarily minds cannot understand those transactions. Take the 2G spectrum scam. [the 2G spectrum scam was a significant scandal in India involving a shortfall of hundreds of millions of dollars in the licenses fees Indian telecom companies were supposed to pay the government for their 2G cellphone spectrum licenses. It is believed that officials were paid off so that lower license fees could be paid]. People still don't know what was the bribe and who paid who the bribe. Nobody knows! There are now 8 or 9 books on the scam and none of these books will tell you who did what. There are lots of charts about where money flowed but what component of that was the bribe and what component of that was legit … why none of these money trails link back to the primary accused … nobody can explain.

[And the big challenge with these high-ticket bribes is that…] The rate of evolution in corruption keeps pace with the rate of innovation in anti-corruption. It will always innovate and modify to protect itself against new approaches.

There are many honest people in government, otherwise this country would be in tatters! If everyone in government was corrupt … but what the system does is almost assume everyone within government is corrupt but that everyone outside of government is innocent. But this is not true at all.

The government represents a sample of the population, so if half the population is corrupt then half the people in government will probably be corrupt, it is not dramatically different. So how does one prevent potential witch hunts or the waste of law enforcement resources following false leads?

With the novelty effect – it feels very difficult to maintain the volume [of complaints] and make grand claims like place "A" is more corrupt than place "B". After some time there will be regional difference and the data will no longer even sound credible to the people involved and its credible will collapse. There is no easy answer this.

So the most effective use of technology in fighting corruption [in India] has been in sting operations.

There you go after high value targets, you don't go after petty corruption at the bottom of the pyramid. You use a competent team, people that know the technology and know the evidentiary value of the technology in a court, you work within the immunity afforded by law to media organizations. Then you build a proper case. Unfortunately the last case that strikes at the heart of corruption in India – the corporate post – you saw how it was crushed by mainstream media because it was not run by a main stream media outfit. And the enemies were too big: the banks that are money laundering. And all banks are money laundering. This is the most fundamental problem to address really because when the bulk of society is outside the tax bracket or the bulk of their transactions are outside the tax bracket then you have normalized criminal behavior in society. It means it is accepted that the average person will engage in tax reduction or tax avoidance. Those type of really high ticket targets – and there are ways to use technology to address issues like this – it needs champions. What "I Paid a Bribe" is trying to do is say that a particular configuration of technology is going to be the solution and that the crowd will address its own problem. This is the assumption there.

That is very different than saying that the solution is never technology, the solution is always people and we need people to drive this, and ethical institutions that are created by these people to drive this. Than what your story of change automatically takes for granted is every time the technology innovates around your current technology fix your going to innovate also. With the other narrative since your technology configuration IS your solution then somehow you hold it sacrosanct and you assume that the corruption is never going to innovate around your fix.

David Eaves: So I think the hope of I Paid a Bribe - right or wrong - was that you could have a radically scaling solution…


If you are going after high-ticket forms of corruption you have to be sophisticated. If you are going after low ticket, bottom of the pyramid forms of corruption you may not have to be.

Take the UID project (biometric ID).

In order to crack bottom of the pyramid corruption, one way to do this is by using biometrics so that the actual recipient who IS supposed to get subsidy or welfare actually gets the subsidy or welfare. But if you just think that through you realize there are so many problems with that.

When the intended beneficiary goes to collect he could be told there is no connectivity, or there is no electricity, or even if the authentication system answered a yes he could be told it was in fact a no. Or even if the authentication system answers a yes he could be given half his entitlement instead of his full entitlement.

And even if the authentication system says yes he may have not been eligible in the first place but could be a local elite who is taking the entitlement.

But on the other hand if we had a system through which there was necessary disclosure of everybody receiving the entitlement and their details were published on a public notice board in the village… then people in the village themselves could tell if those people were the right people to get these subsidies. The local media could come and inspect this, civil society organizations could inspect this. So it is the government becoming more transparent to the citizen rather than the other way around, which with the UID is what happens – the citizen becomes more transparent to the state.

Technology has to be configured in a very careful way at each stage in order to address different types of corruption, and each of those specific technologic choices you make can either increase corruption or decrease corruption or increase the power asymmetry or make it more equitable.

Eaves:So what are the conditions in which I Paid a Bribe might work more effectively?


It might be that I Paid a Bribe only allows for one type of participation. If you look at other forms of commons based peer production projects. I think they offer many types of contributions to the project. So if through the examination of a particular government budget a platform identifies all the public works that are going to be constructed using that budget and then allows state or town or even a locality to monitor the progress on those works then it allows for contributions that are necessarily antagonistic but it also allows for contributions that are antagonistic. So it allows for a conversation with a much more diverse set of people participating.

With I Paid a Bribe, because they want the scale they have taken the simplification very seriously. And it is a light touch transaction – which means once you register a bribe you can more or less forget about it. I think other peer production projects, even Wikipedia, once you start editing a particular page, you may develop an interest in that page. You can set a notification system that will notify you every time that page is edited. It is like the same story on a mailing list, that the rules that existed during the initial days when the mailing list was very small change and get more sophisticated as the mailing list scales. The technological solution has to grow to compliment the age of the community and the community has to be gardened very carefully to retain members. I Paid a Bribe risk ending up being a venting mechanism for the post bribe trauma moment. Which is fine! That is useful in of itself. But if your vision is more than that, if you want to have a movement which is an anti-corruption movement then it cannot be just be a venting mechanism for the anti-bribe trauma.

This is the complication. I'm not dismissing I Paid a Bribe. All I'm saying is, in order to build a community and take it along with you, you cannot keep anything sacred. Not your technological configuration, not how you message, address and bring people into the community, not even your target community – that might change – so it is just like the challenge we have in growing Wikipedia in India. If you were that simple we would the same solution for the different language wikipedias and we would do one homogenous thing across the nation. But the Kannada Wikipedia has only seven active editors, as does the Punjabi Wikipedia, the Hindi Wikipedia has hundreds of editors but they don't like each other and they don't meet in real life. Very unlike the Maleren Wikipedia. So the strategy that needs to be employed is very different for each of these communities. There is almost an automatic tension between simple, scalable and authentic bottom up movement forces.

So as the number of bribes reported on I Paid a Bribe increase – what do you do to ensure that the quality of those reports gets better? And how do you move people from being anonymous reporters to the system to becoming more and more identified? How does a senior cohort of the community get created? I don't know – these things may exist but I don't see it from the outside.

Look at the different in other peer production projects. The brand of the individual and the attribution afforded to each individual is quite explicit and public. I Paid a Bribe is mostly anonymous although I think some of them choose to be public.

Eaves: And you need to feel incredibly confident in your personal security and status to be public in I Paid a Bribe.



This post has been corrected to fix a transcription error. Sunil Abraham said the Kannada Wikipedia has only seven active editors.