AI and the human factors

The Digital Ape: How to live (in peace) with smart machines
Nigel Shadbolt and Roger Hampson
Amaryllis
347 pages; Rs 499

The proliferation of artificial intelligence (AI) — from Deep Blue and Watson to Siri and Alexa — is presented as a celebratory as well as cautionary theme for our age. Even as technology overtakes more and more routine tasks, tech leaders like Elon Musk and Stephen Hawking warn us that the final destination of this journey will spell the end of mankind.

 
In the book under review, Nigel Shadbolt and Roger Hampson, both associated with the UK’s Open Data Institute, take a largely sanguine view on the issue. This is a history of the growth of AI, from its beginnings in the theories of Alan Turing to now, when it is the backbone of the larger web of technology that surrounds us. We may not have driverless cars yet, but the smartphones in our pockets are no less a triumph of AI.

 
Dismissing fears stoked by the likes of Musk and Hawking, the authors argue that AI, for the foreseeable future, is unlikely to reach a stage where a robot uprising will defeat humans in a large-scale wipeout. They rightly dub such scenarios outlandish, spurred more by the fertile imaginations of science fiction enthusiasts than any real-world evidence, where smart machines are more likely to be lawn mower discs than murderous, self-aware beings.
 
While reasonable, this evidentiary approach assumes that there is no immediate risk to humans from AI. In other words, what the future holds is up in the air. This supposes that technology will some day come to mirror natural intelligence to such an extent that it will tip over into consciousness, an evolutionary game in which the deeper you go, the closer you get to becoming human.

 
The authors present recent advances in AI to bolster this theory. The much-celebrated victory of AlphaGo against the world’s best Go player is cited as an example of how machine learning — in which software becomes better by observing millions of patterns — is a precursor to other, hitherto-unexplored vistas that may end with the sparking of something lifelike in those inanimate machines.
 
That, however, is unlikely to happen. A century-and-a-half after Charles Darwin published his seminal tract on evolution, we are no closer to understanding how a bunch of carbon molecules can yield the sort of intelligence that, say, produces sublime music. It is hard, then, not to spot a glimmer of scientism in the authors’ cautious optimism about a future of conscious robots. It may not be au courant to bring up God in a discussion of high technology but His absence makes Him conspicuous here.

 
The book’s look into how current technology is already giving us a glimpse into the future of AI is more interesting. Why, the authors ask, do we need robots to be anthropomorphic? Is it easier to relate to a vacuum cleaner if it looks less like a machine than a human? And if we get there, what are we going to do about robot rights?
 
As earlier, the authors take a scientific view of the issue, elaborating how researchers are working to incorporate neural network architecture into new-age robots. The discussion focuses on the Turing test, a means to figure out whether the entity one is interacting with is human or artificial. But even the best technology is a slim substitute for human connection. An excited look into how digital assistants can mitigate loneliness among the old accentuates this feeling.

 
As the 2013 Hollywood film Her demonstrates, robotic consciousness if it comes to fruition, is likely to branch into all manner of directions, not all of which a human would grasp. Samantha, the voice of AI in Her, starts as a personal assistant to the film’s protagonist, Theodore, who gradually falls in love with her.
 
But there is no Samantha — or, more precisely, there is no woman called Samantha, merely a software — a fact that Theodore comes upon brutally when she tells him that she has a relationship with hundreds of thousands of other men just as she has one with Theodore. Her tracks the devastation this unleashes on Theodore.

 
The Digital Ape skirts such interesting asides into the nature of technology to present a largely unidimensional view of what AI will be to us, humans. A chapter on the future of work reiterates the long-standing argument about how new technologies replace old jobs to create new ones. Towards the end, the book undertakes a discussion of politics — Donald Trump, Brexit, and post-truth — with a critique of social media that lends little weight to the book’s subject matter.
 
All said, AI, the next big frontier in the ongoing saga of technological leaps, may make life easier but it is unlikely to provide answers to fundamental human concerns. The authors’ eager evangelism is focused too much on efficiency gains at the cost of more substantive concerns.