AI can deepen democratic governance. Is the state ready?

There has been a lot of talk about what Artificial Intelligence will do to the economy. In particular, there has been concern that it will eliminate many well-paying jobs as companies adopt AI more. There has also been some coverage of how AI is increasing productivity in other sectors – generative diagnostics in healthcare, predictively balancing vast electricity grids, forecasting wildfires, etc.

Market and society are seeing their own adoption rates of the new technologies, but what about the state? While there has been some use of AI in public administration, we have barely scratched the surface. The number of AI-driven sarkari applications will grow, but why stop with administration? Why not ask: What can be done with emerging technologies for democratic governance itself?

This is a bus we have largely not boarded. Even as technology became ubiquitous, governments adopted it slowly, often doing the same old things digitally. One obstacle was limited knowledge of emerging capabilities; governments are poor at training their people. But a second, bigger hurdle was plain unwillingness. The status quo suited those in the state. It even suited those in the opposition, for one day they might be in power!

When will democracies bound across this steeplechase? One doesn’t have to be a soothsayer to imagine that things will change, somewhere, sometime. To understand why, let’s look back a little.

Around the world, democratic governments have expanded enormously, doing far more than they did in the past. Leaders, once in power, widen the scope of state intervention. But representative democracy was not meant to be a battleground for competing agendas. It was to enable people to influence the processes and rules governing them.

Today, technology has removed many of the early hurdles that necessitated representation. It is far easier for people to directly vote on legislation tabled in the state Assembly or even Parliament. If our representative is only pressing a button on our behalf, surely that can be done by us in our homes.

This is particularly true when those making decisions are not even reading the material needed to make them. So many bills are passed in representative bodies with hardly any discussion. These cannot be called ‘representative’ action. If my MLA has no time or inclination to read a draft Bill, why not let me do it myself?

Many of the standard answers – not everyone can make informed decisions, not everyone has time, etc. – are jaded. Some do have the time, inclination, and ability to read, understand, and decide. And if people were able to make more choices themselves, they might start making the time and having the inclination to do so.

This may seem far-fetched on the scale of a nation or a state, but it’s not that difficult to imagine if we’re voting on how local ward budgets should be spent. People in Switzerland have been voting on micro-issues: things like changing a bus route. At that scale, a lot is doable. Perhaps that’s the way things will change; slowly and locally at first, then in slightly larger settings, then suddenly.

One problem that AI has reduced greatly is the asymmetry of information between the state and the people. Governments hold a lot of information, and citizens don’t know even one per cent of it. But there is an irony here: those in government themselves don’t know much about what is happening, because each department withholds information from others, or they don’t proactively share it. Intelligent agents can help. Around the world, governments are opening their datasets and inviting the public to help tackle problems. Bridging the democratic distance between voters and decision-makers will gather pace.

Tinkering with this is exciting. Earlier this year, after my visit to Kalaburagi, I was keen to figure out what can be done to develop a backward district like that, and what state and national governments are doing. Working with Sathya Sankaran – well known as our city’s Bicycle Mayor – I discovered that Claude could compile, categorise, and present this information in less than a few hours for a few thousand rupees.

I’ve shown this work to District Collectors and ministers in charge of districts, and they see its potential. Another version of the same tool can be built for cities. The same Claude brain turned the budget speech of the Bengaluru South Corporation into a trackable platform for the promises listed in it. For as little as a few hundred rupees of tokens! Making something for a local ward should be – and may one day be – a middle school project.

Government should be for the people and of the people, but the part that holds the most potential is the ‘by the people’. And it is getting a major boost from AI.

The writer is a social entrepreneur, founder of Mapunity and LVBL, and co-founder, Lithium, wakes up with hope for the city and society, goes to bed with a sigh.

Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.