By Alison Tilley
Open data is like motherhood and apple pie. Everybody thinks it’s a good idea. I am suspicious when everyone is in favour of something. Especially when they have conflicting views of how the world works, there is something going on there which needs consideration.
So when governments, and business, and social justice advocates are all in favour of open data, what is going on? Well, allegedly, a revolution. The information that exists in the world is growing exponentially, and with it the possibilities of profound change. We all want change, we say – to change is to improve, and to perfect is to change often.
But what do we all want? Are our goals the same? If they are, that seems so unlikely as to be incredible. Business usually argues for stability. You can make money in any situation, they say, provided you can predict what will happen. Yet here business is apparently in favour of big data, the internet of things, disruption. Governments don’t generally want change. They want voters to vote for them, and tweak systems to try and get votes, but the only way governments change profoundly is when they are overthrown, and even then the new state often looks strangely like the old. And yet many seem to be in the same corner as social justice advocates, arguing for open data.
Classical neo liberals thinks open data delivers information symmetry, and thus better functioning markets. A function of the explosion in computer processing power is the information that is available and that we think we can see patterns in. That’s money right there – information asymmetry, which companies can exploit. So open data gives us both symmetry and asymmetry?
If it gives us symmetry, then that theoretically gives us more equal power. That is the classical ‘information is power’ argument. If I know what you know, we are equal in that respect, and if you are the state and I am a citizen then I am more powerful than I was, and I can hold you accountable. Or so the theory goes. So good governance advocates are in favor of open data. Governments are also in favour of open data, because they say it helps deliver better services for less money.
Politics makes for strange bedfellows, they say, and it may be indeed that we are all in favor of open data. But perhaps before we conclude that too quickly, let us consider a little more closely.
We probably agree on a definition of open data.
The question of what open data delivers is perhaps something we do not agree on. That is because what open data delivers in theory is quite different to what it delivers in practice. What data is open to whom? Open data evangelists would say, default open – that is information is made automatically available, as a matter of law, policy and practice, as open data. Only very limited categories of information are not available.
Well yes, but whose information? Government’s? Or big business? Does Google have to make its data open by default? Or Apple? No, wait, says business. Trade secrets, copyright, research information, that is our business model. We can’t operate if we don’t keep secrets.
Government information? Well yes, says government. But not national security data, or data which relates to international relations, or tax or, even in the case of the Chinese, budgets. Those should not be open data. We need those secrets.
Even data evangelists are a bit more wary of open data now. In a post Snowden era, the notion of a state which can collect data on you anywhere at any time, and apparently does, is suddenly not so attractive. So there are the privacy concerns, and the issues around confidentiality. Ontario’s privacy commissioner discovered last year that the mental-health information of some Canadians is accessible to the FBI and U.S. Customs and Border Patrol.
Some Ontario police services routinely uploaded attempted suicide calls to the Canadian Police Information Centre (CPIC), to which U.S. border guards and the FBI have access. Not so keen on open data now, are we? We all need secrets.
So the devil is in the detail. My open data is your classified data – your private data is my business model. So, like applehood and mother pie, we agree. Until we get down to what is available to who and when.
Release crime data, by police station, every month, says the Khayelitsha Commission of Enquiry in Cape Town, in South Africa. That will help build trust in the police. Once a year, says the South African Police Service, and never at police station level, only in aggregated form. Otherwise criminals will use the information for intelligence purposes. Yes, that is really their argument.
We want land audits of who owns what land, say land activists. No, says the state – we aren’t going to give you the information that would trigger land invasions. Information about the nuclear power framework agreements South Africa has entered into with Russia? Hell, no, says the Department of Energy. The release of the framework agreement would compromise “sensitive negotiations”. The framework agreement was found to be available in Russian on a web. So whose sensitivities are we compromising?
So what can you have? The budget, says the government. What about the bids put in by different departments to Treasury, so you can compare what a department asked for, and what was allocated? Well, no. Not that.
Can I have the right to be forgotten? Can I require that search engines delete my data? Yes, in some jurisdictions. In others, that’s a violation of the right to a free press, and freedom of speech.
Suddenly the practice as to who will give what to whom and when looks very different from the words that we all agree on. That does not begin to deal with the questions of what we understand all this data to mean in the first place.
The phrase, ‘there is no such thing as raw data, but only cooked data’, is apposite. You can, as Code4SA did, map all the marriages in South Africa, and find some very interesting patterns. You could potentially even draw legislative, policy, business and advocacy decisions from that. But that’s only marriages between men and women, registered, and conducted by a marriage officer. No civil unions, customary marriages, domestic partnerships, Muslim marriages, Hindu marriages, or a myriad of other relationships that don’t tick that box. So we have a fundamental agreement about what constitutes a marriage? Clearly not. So when we have data that tells us about marriage, we must understand that it ‘cooked’ data, not ‘raw’ data. It assumes a range of parameters about what marriage is, which are not parameters that are even legally sound. Let alone culturally agreed.
So what does open data really mean? Whose data? When? How? These are some of the issues that we need to get to grips with if we are to do more than agree on generalities. A false consensus may give an appearance of accord, but a false accord is not a basis for a revolution.