Machines and robots rule the world but, can they ever take the place of humans?

By Vanessa Barreno

Many will remember the Japanese-fuelled advances in micro-circuitry and semi-conductors throughout the 1970s and 80s, leading to explosive growth in electronics, computers and consumer gadgets. That was then. Now we are under no illusion that everything can be – and is – done by machine. We ask the ultimate question: Will humans gradually be replaced by machines or are there things machines just cannot do?

Back to the future

First came bi-polar transistors and digital integrated circuits, followed by Very-Large-Scale Integration (VLSI) in the mid-70s, consisting of thousands of on-off switches on a single chip. Microcomputers, medical equipment, video cameras and communication satellites are common devices made possible by integrated circuits. Since then, parking, airline check-in, travelling on the metro, factory production lines and DIY check-outs in supermarkets have succumbed to mechanisation.

In 2008, “Mechanization of Modern Society” was presented by the Humanities Center at Carnegie Mellon University, Pittsburg, US, during the Carnegie Mellon International Film Festival. Countries as diverse as Argentina and Austria raised topics ranging from the processes involved in the mass production of fast food and clothing to parents’ efforts to take advantage of scientific advances to “design” the perfect child. The award-winning, thought-provoking films raised important questions about technology and modern society, epitomised by La Antena, a black-and-white silent film in which citizens of a future society live in a voiceless world thoroughly controlled by television.

Chess champ beats computers; hands down

25 years ago in Hamburg, Germany, World Chess Master Garry Kasparov played in a simultaneous exhibition against thirty-two different chess computers, one machine to the next, making his moves over more than five hours. Four leading chess computer manufacturers had sent their top models as his opponents, including eight named after Kasparov, so the stakes could not have been higher.

Kasparov won hands down, thrashing the machines 32-0. The challenge was all the greater given a sticky moment with one of the manufacturer’s doppelgänger ‘Kasparov’ brands. This made him intensify his efforts – if the machine had won or drawn, “people would be quick to say I had thrown the game to get PR for the company,” he explained. “Eventually I found a way to trick the machine with a sacrifice it should have refused. From the human perspective, or at least from my perspective, those were the good old days of man vs. machine chess.”

Fingertip failures

According to the Wall Street Journal, the New York Stock Exchange has been promoting its hybrid model of trading as the best of both worlds in which humans and computers trade side by side. Yet at the height of global financial meltdown – when the model should have functioned at its most efficient with human lives depending on it – it failed.

The Exchange shut down computer trading of some stocks, handing them over to human traders on the floor for minutes at a time to slow the market, as humans help find the right price for volatile stocks.

But as NYSE is the only major exchange using this system, the rest of the market’s computerised exchanges continued trading, building up huge volumes of sellers with virtually no buyers. Instead of steadying the volatile market, the NYSE floor traders were unable to cope with the volumes.

Fingers were pointed as to who was to blame for the “tumultuous trading day”. Rival exchanges claim NYSE’s machine-to-man system made matters worse because it removed a key place to trade during those crucial minute intervals, discouraging some buyers who might have been able to stabilise prices.

Machine vs. “human-engineered” translation

Where hardworking translators in all manner of situations innocently muddle millions of vital “human-engineered” interactions, machine translations are increasingly employed despite their often comical lack of reliance, and they are often no worse than human errors.

The efficiency of machine over human translation was demonstrated in the midst of Haiti’s earthquake in January. Well-meaning aid teams flooded in speaking dozens of languages – but not Haitian Creole, the country’s second official language along with French – so a trapped survivor with a mobile phone would have been unable to communicate with the rescuers.

Within three days, forward-thinkers at the Language Technologies Institute of Carnegie Mellon University in Pittsburgh and a network of volunteer developers produced a Haitian Creole speech and text data machine translation system. Though not faultless, it worked and having collected the data, Carnegie released it publicly with minimal license restrictions to allow others to develop Haiti’s language technology.

And even though ubiquitous Google has entered the machine document translation market, human translations have to come first for Google Translate to work.

Humans vs. machines in Barcelona

Research teams at the Human-Computer Interaction Department and Computer Vision Centre of Universitat Autònoma de Barcelona have developed a cognitive computational system consisting of video cameras and software able to recognise and predict human behaviour, as well as describe it in natural language. Codenamed HERMES – Human Expressive Graphic Representation of Motion and their Evaluation in Sequences – its numerous applications include intelligent surveillance, protection of accidents, marketing and psychology.

HERMES analyses human behaviour using video sequences captured at different focus levels and distances, studying body posture and the individual’s face to give a detailed picture of facial expressions. This information is processed by com- puter vision and artificial intelli-gence algorithms that permit the system to learn and recognise movement patterns.

This ground-breaking development offers important innovations through the natural language movement captured by cameras, and the possibility to analyse and discover potentially unusual behaviour based on the movements it recognises to give off warning signals. As a practical application, HERMES could send a signal to a metro station control centre after capturing the image of someone trying to cross the tracks, or alert a medical centre if an elderly person living alone suffers a fall.

The irony of machines monitoring humans for our safety and survival has not gone unnoticed. While electronics have a yet bigger part to play in medical advancements and clearly offer life expectancy and wider social benefits, doubts remain as to our collective over-reliance on machines’ capacities to ever completely substitute us as humans. As TIME Magazine once highlighted in its piece on ‘What the Fancy Machines Can and Can’t Do’: “The machines we have to analyse the body are complex. But not nearly as complex, or as wondrously made and maintained, as the human body itself.”