Viswanathan K's Thoughts

Is AI over-hyped? Or is it going to transform the landscape so drastically that it would be unrecognizable? Like how electricity did a century back?   If it will/does shake up the job market drastically, how quickly/slowly would that happen? At the pace of electricity (quite fast, but nowhere close to overnight)? Or much slower? Or way, way faster?   It was in the context of these questions that (right or wrong) China’s approach being so different from the US is worth checking out.   AI, if it were to be as transformative as some say (fear?), would cause massive job losses and social upheaval, the backdrop to every revolution. Which is why the Chinese government (single-party rule system) is wary. On the other hand, China can’t ignore AI, given how much potential it has, plus the risk of its arch-rival running too far ahead. Therein lies China’s AI dilemma. What then is China’s AI approach?   Since 2023, all public facing AI models must be filed with...

The government is terrible at many things. But it has also done other things extremely well. Like polio eradication and conducting elections on massive scales. Is there any pattern? Is the government good at certain types of activities? Which ones? And can it avoid the other kinds?   Pritchett and Woolcock came up with a model to analyze just those questions. It involves splitting proposed activities into a 2 x 2 grid – one axis is the action (Discretionary or Non-Discretionary); the other axis is the number of transactions involved (Huge aka Intensive or Few aka Non-Intensive):   The bottom left corner ( Policies ) is about actions that require discretion but are not done often. Examples include changing tax rates or setting eligibility criteria or location of dams. These are choices and decisions that have to be made, but since they aren’t done often, in theory , a government could use the relevant experts to make the right decision most of the time. But like any disc...

In Missing in Action , Pranay Kotasthane talks about eight ways of thinking that don’t make for good policy making. First up is the idea that you know what a good policy looks like. Nobody does, because the world is too complicated. Actions have unintended consequences. People don’t always react in ways you (or the policy maker) hoped. Second one is related to this – a belief that one needs to be consistent on policy matters. No, he says, one should be flexible and change one’s views based on how policies have fared. Un-learning is just as important as learning.   Third is the belief that good intentions translate into good policies. The fate of all attempts at prohibition across the world are a perfect example of this. Fourth is the wrong idea that a policy is good, just that its implementation or execution is where the fault lay. No, he argues, a “policy formulated bereft of implementation details” cannot be a good policy. All of us know both these points from our own lif...

Sometimes, indulging your child’s whims can set off a change in policies, writes Montek Singh Ahluwalia in Backstage . One such example was the establishment of Maruti Udyog Limited (MUL). It was established to realize Sanjay Gandhi’s dream of producing an affordable people’s car. The Planning Commission and the Finance Ministry opposed it – cars were a luxury item, why spend on that? Indira Gandhi approved nonetheless. “(Unintentionally, MUL) helped change attitudes on many important aspects of industrial policy.”   First, it forced us to accept our limits. Yes, we could manufacture a car, but we were nowhere ready to design one. The hunt for a collaborator began. “This also opened the door to a more relaxed approach to the import of foreign technology in other sectors.” Why Suzuki was selected is amusing. No other Japanese car manufacturer was interested. Suzuki though was open to talks since they were still a motorcycle manufacturer and had just introduced a small car...

In the last part of his book Privacy 3.0 , Rahul Matthan presents his view on how the laws on privacy should be framed in India. He points out that the Aadhar horse has been out of the stables for a while now, and has been unifying various databases – from PAN to bank accounts to your mobile number. And it has undoubtedly yielded benefits to all – the UPI system works only because the banks and phone numbers could be connected via your Aadhar ID. The eKYC that Aadhar has enabled cut down the cost of verification from ₹ 1,000 to ₹ 60. In turn, that has reduced the costs of the lending sector, which then opened up the market for low value loans, from ₹ 25,000 onwards to become viable. The potential benefits in the healthcare industry via a system like Aadhar are enormous – one could identify which areas are prone to which diseases; or correlate symptoms to diseases in ways no individual doctor can.   That acknowledged, he points out that in the digital age, more and more comp...

Given the future we are headed towards with its AI, self-driven cars and drones (and who knows what other forms of “smart” software”?), governments world over worry about security implications. And yes, governments can look outside the bureaucratic box, as seen in Bruce Schneier’s description of one such US initiative a decade back: “The Department of Homeland Security hired a bunch of science fiction writers to come in for a day and think of ways terrorists could attack America. If our inability to prevent 9/11 marked a failure of imagination, as some said at the time, then who better than science fiction writers to inject a little imagination into counterterrorism planning?” Schneier though sees a problem with consulting sci-fi writers: “More imagination leads to more movie-plot threats -- which contributes to overall fear and overestimation of the risks. And that doesn't help keep us safe at all.” Rather, he prefers this model: “Science fiction writers are c...