A data apocalypse is coming our way, with massive identity theft unleashed on a less-than-suspecting, sometimes illiterate populace.
Western Cartesian “Science” strives to divide a system into smaller and smaller pieces, assuming that the ultimate truth is in the pieces. But that’s not true.
As a technology strategist, I am often faced with a dilemma: there are many new technologies that have compelling value propositions, and there is a natural tendency to be excited about them. In these pages, I too have written about new tech—Machine Learning, Blockchain and Genetic Engineering (May 2016)—and its potential impact. Indeed, I have recently been looking at how blockchains may make transactions friction-free and intermediary-free.
But I am impelled to also look at the other side of the picture—what might be the downside of rushing pell-mell into new technologies that may have big downsides that we are not aware of? The principle of unintended consequences, if you will. This makes me a techno-sceptic. Some even call me, cruelly, a neo-Luddite.
Perhaps I am, but maybe the sentiment is more akin to the views of Chinese essayist Lin Yutang in The Pleasures of a Non-conformist. There is value, and indeed pleasure, in being a sceptic and a non-conformist. It is only human to be seduced by the new stuff, because it is so full of promise, and we are, I suppose, programmed to be optimists, but someone has to question the emperor’s new clothes.
Often the hype gets far ahead of reality (the Gartner Group’s Hype Cycle is an attempt to fit irrational exuberance and the consequent trough of disillusionment into a pattern). The cognoscenti become evangelists, and they cannot stand it if anybody—even a relatively neutral person—questions their facile axioms. But questioned these axioms must be, and the insiders often become quite bigoted and nasty, perhaps because they are industry employees with an obvious axe to grind.
I spent most of my career in the computer industry, much of that in Silicon Valley, but I don’t feel that industry is sacrosanct.
It makes me nervous that some insiders are unthinking cheerleaders for
a) Western ideas,
b) “science” (although most of the time they mean technology).
I recently spoke with a public health expert whom I think well of, but he airily dismissed those “unscientific” (translation: non-allopathic) practices.
And I listened to a chat with a Nobel Prize winner regarding the alleged wonders of “telomers”. I was thinking, “snake-oil du jour”. The so-called medical “science” is rife with fraud, bad science, and the simple expedient of conflating correlation with causation. Case in point: the cholesterol scare, which, it appears, was manufactured by a few Harvard researchers in cahoots with Big Food. The US Government is set to withdraw longstanding warnings about cholesterol’” from the Washington Post in 2015.
No wonder it resonates when statistician and philosopher Nicholas Taleb of Black Swan fame roundly condemns some “science” for unverifiable results, drugs that do not work, psychology studies that cannot be replicated, and forecasters whose models are good at anything but predicting behaviour.
Sometimes I think “science” has become a shibboleth to hide incompetence. Western Cartesian “Science” strives to divide a system into smaller and smaller pieces, assuming that the ultimate truth is in the pieces. But that’s not true: they cannot understand or explain “emergent” behaviour, such as the apparent intelligence of a beehive or an anthill. Nor, for that matter, can the West define what Life itself is: it is surely much more than the $83 worth of chemicals that constitute a human body.
There are also many western inventions that turn out to have unfortunate consequences even if they are convenient. The flush toilet is nice, but we can’t use all our water for that purpose; the car is nice, but it has also led to global warming; plastic is useful, but just look at every roadside or railway line in India, or the Texas-sized gyre full of plastic in the Pacific Ocean.
At the moment, there are three areas in technology that concern me: genetically modified organisms, nuclear energy, and financial technology (fintech). The first two because they are in the news for very negative reasons, and the third because of the extraordinarily inflated expectations of the good it can do, without due attention to the potential problems.
GMOs have been on my mind partly because of an ongoing debate on Swarajya between proponents and a dissenter. I wrote in support of the dissenter, about the fact that Indian cattle breeds are on the path to extinction despite their superior full-life-cycle milk delivery and their apparent resistance to mad cow disease. Taurus indicus also yield A2 milk, which is superior to A1 milk provided by imported Taurus taurus (eg. Holsteins) cattle. I just learned from my diabetologist friend that A1 milk is suspected in the increasing prevalence of Type 1 diabetes in the West.
There are, however, not only health concerns with GMOs. The very business model is one of customer lock-in, that is, seed companies force farmers to be forever dependent on their GMOs. Furthermore, the very virtues that the GMO organisms are celebrated for—such as their resistance to certain pests—may well be transient. If I am not mistaken, the recent news that Monsanto has been told to withdraw its Bt cotton from India is based on the fact that in fact their product no longer is immune to the bollworm, which was its alleged virtue.
The problem with the business model, of course, is that there is the likelihood that the GMO organisms (as with other imports such as the Taurus taurus breeds) will displace native species, in addition to forcing the farmer to buy seed every year from the Monsantos of the world. The latter problem (typical of customer lock-in, as in the Delta Model articulated by Arnoldo Hax of MIT) means oligopolistic profits, usually not good for the consumer.
But the former problem is more insidious. The native species have typically evolved to be in rough equilibrium with the local climate, pests, soil, temperature, and so on. The imported species typically will have to be nurtured, as it were, in a hothouse or controlled environment, unable to take on the vagaries of local conditions that native species would not be fazed by.
Besides, the dangers of monocrops persist. Despite there being many native species of banana, there is a global tendency to converge on a single variety, Cavendish or Robusta. A single catastrophic pest can wipe out the entire crop of this monocrop.
We mess with agriculture at our peril. Opening a Pandora’s Box with unforeseeable consequences (not only known unknowns, but unknown unknowns) is a high-risk strategy. What if the sceptics are right, and GMOs do cause catastrophic crop failure? “Oops, we were wrong” is not a good enough answer in that case.
A pragmatist would want quantifiable risk, to be able to make an informed judgment about whether to proceed with the programme. A recent World Economic Forum article “How human error could have created the Sahara desert” is a case in point. It appears that bad agricultural, hydrological and animal husbandry practices may have created a tipping point to degrade the once-fertile area into a desert. Also see the dire warnings in “Climate-driven species on the move are changing (almost) everything”: global warming is apparently changing the normal range of land and sea-based species so rapidly we could see ecological collapse.
The second sector that excites furious argument is nuclear power. There are well-meaning proponents of this technology in India, who intone gravely that lack of power will be a huge problem for India (which is true) and that nuclear power is the best way to solve this (which is not true).
I have just one word for them: Fukushima. They expect to spend $188 billion on the cleanup, and are nowhere near solving the problem. In the core of crippled Reactor II, the radiation levels are so high that robots sent in for reconnaissance are fried in minutes. The level is as high as 530 sieverts/hour. The total radiation that a human can safely be exposed to in his lifetime—yes, entire lifetime—is 1 sievert.
If they want another word, here it is: Hanford. The former nuclear site is expected to take $120 billion and 40 years for cleanup. Admittedly, Hanford is a weapons site, not a civilian nuclear reactor, but that gives you an idea of the cost of cleaning up nuclear waste. This is over and above the potential costs in health and contaminated land and groundwater, and also ignores the ever-present danger of dirty bombs in the hands of terrorists.
So there definitely is a business model problem: even if nuclear energy is relatively cheap in terms of operating costs, the capital costs are high, and the total cost of ownership (TCO) is astronomically high. This cost-overrun problem has just about fatally wounded Japanese technology giant Toshiba because its nuclear subsidiary Westinghouse is driving it into bankruptcy. Areva, another nuclear firm, is in serious trouble, too.
Meanwhile, India appears to be moving ahead with its plans for nuclear reactors including some from the selfsame Westinghouse and Areva. Pardon me if I am reminded of other instances where dangerous Western technology has been dumped on Indian dupes: remember chloromycetin, an antibiotic banned in the US, which was happily prescribed by Indian doctors in the 1970s?
In both these cases, Indians have been so blinded by motivated propaganda and hard sells by snake oil types that we have committed large amounts of money. In the third case, we are on the point of committing a different mistake: fixing a piece of the system while the thing in its entirety needs to be rethought and re-architected. This is the fallacy in computer science referred to by Donald Knuth when he wrote “premature optimisation is the root of all evil”.
It so happens that the move towards electronic banking, JAM (Jan Dhan, Aadhaar, Mobile), and perhaps even blockchains is afflicted by this lack of a system-wide perspective. I have been reading a number of articles that attack Aadhaar from several perspectives: loss of privacy, accidental or deliberate leakage of personal, private data, and so on.
Nandan Nilekani made a spirited defence of Aadhaar, which boils down to an assertion that nobody has actually breached the security of the database that maintains all our fingerprints and iris scans. He may well be right, in the narrow sense, that the database is (at least so far) unhacked. But that doesn’t matter. The point is that Aadhaar is a 21st-century authentication mechanism grafted on to a 19th-century system of forms in triplicates and other endless, and meaningless, pieces of paper.
I encountered this personally when I went to buy a Jio Sim. They promised me that the process would be painless and I’d have my sim activated very quickly if I produced my Aadhaar card and had my fingerprint scanned. And so it turned out to be: within a few seconds, my entire information, including address and even my extraordinarily bad photograph, appeared on the retailer’s screen. In theory, he could easily sell all my personal information to anybody who wanted it. And furthermore, he could capture my fingerprint information by using a man-in-the-middle tactic. And sell it.
But the relevant question is: “Why is Aadhaar populating a set of old printed forms?” If its role is merely to authenticate people, all that the Jio guy needed to know was: “Is this person authenticated to be who he claims?”. A yes/no answer is all he needed.
Yes, we understand that the government (like all other governments) wants to know who is buying Sims, for obvious security reasons. But for that, the question is only a yes/no question. That whole issue of “need to know” is at play here. The Jio guy did not “need to know” anything more about me.
Thus, the integrity of Aadhaar is not the problem; it is the fact that the process is not thought-through and rationalised. There was a similar problem earlier with pre-paid recharges, where the retailer was required to write down the names of the people who came to do so. It turns out, at least in Kerala, that retailers were selling the names and numbers of attractive young women, who would then be stalked by young thugs.
I personally experienced this, for a young man mistakenly kept sending SMSs to my number, and I did nothing, merely observing his messages. It was evident that he didn’t know the girl he was wooing: he kept inviting her to come down from the apartment she lived in so they could talk. He also sent small recharges to my phone, so ‘she’ could send him SMS messages. This went on for a while until he figured out that he had the wrong number. So a system set up with good intentions ended up preying on young women.
Similarly, I am concerned about the security of blockchains too. It is likely that the hashed values themselves, stored in P2P fashion across thousands of servers, will be indelible. But I am not so sure what will happen when the hardware security of nodes, especially IoT nodes such as Point-of-Sales devices, may not be guaranteed. Will this mean the very entry of transactions will be tampered with?
Data in flight is indeed vulnerable. A few months ago, Rumanian hackers attacked an State Bank of India ATM in Thiruvananthapuram without physically tampering with the device, and were able to syphon off the debit cards of a number of customers, causing heavy damage. And Point of Sale(POS) devices are notably less secure than ATMs. I can see a data apocalypse coming our way, with massive identity theft unleashed on a less-than-suspecting, sometimes illiterate populace.
I don’t mean to rain on anybody’s parade. And I am all for a little good-old-fashioned rah-rah. But when this gets out of hand, and nobody is thinking about all those cyber-criminals with keen jugaad minds, then I begin to worry. Hence my techno-scepticism.