modern.fakir
Chief Minister (5k+ posts)
Indian IT's AI: More artificial, less intelligent
Pankaj Mishra,ET Bureau | Jul 13, 2015, 01.02 PM IST
Over past few months, TCS, Infosys and Wipro have been pitching their versions of AI solutions as the next big thing in tech.
Indian IT's AI: More artificial, less intelligent
Over past few months, TCS, Infosys and Wipro have been pitching their versions of AI solutions as the next big thing in tech.
Depending who you ask in India's over $146-billion software industry (and that includes top executives at the biggest companies) the answer can be anything from automating repetitive tasks for clients, to applying predictive analysis for ensuring computer infrastructure doesn't crash down and so on.
Last month, India's largest software company Tata Consultancy Services launched Ignio, its new AI platform. The company talked about how the algorithms that form "the neural brain" of Ignio is capable of performing repetitive tasks. It can also predict peak demand for computer servers and ration storage capacity accordingly, in advance.
Wipro, the country's third largest software company is touting its cognitive computing system, Holmes, as "the Watson killer", referring to IBM's Jeopardy-winning supercomputer that's now being used by clients as an AI solution.
Executives at Infosys, India's second largest software firm, too are using the term AI while describing their current software platforms, most of which are focused on solving problems that occur repetitively.
If John McCarthy, or an Alan Turing were to hear these definitions on what AI can do, they wouldn't be too pleased. McCarthy, a Stanford scientist who originally coined the term "artificial intelligence" way back in 1955, defined it as "an artefact that could actually replace human level intelligence", clearly distancing it from machine learning.
A true AI solution has to be the closest to how a human brain thinks. And it will need to pass "the Turing test" wherein a machine exhibits the same intelligent behaviour as a human when faced with a problem.
As a top scientist advising companies on leveraging AI tools described, an artificial intelligence solution should be able to tackle problems not encountered before.
Over past few months, TCS, Infosys and Wipro have been pitching their versions of AI solutions as the next big thing in tech. These companies are using terms like "neural networks", "mimicking the human brain" and "applying human-like intelligence" in solving the tech world's biggest problems.
Most of the AI applications being touted by India's top software companies are based on using insights from accumulated data to take decisions. These solutions sift through massive chunks of data, identify repetitive patterns, and ensure that the systems automate these tasks.
This is hardly AI.
Let's take the example of the famous Hudson river plane crash in January 2009 when its pilot Sully Sullenberger landed the plane in the water. If he had listened to air traffic controllers, or relied completely on the available data for deciding whether to return to LaGuardia airport or land in New Jersey, he would never have landed the plane in the Hudson river.
As it turned out later, his decision to not rely on automated systems and instead go by his own judgement saved 155 lives.
"Most of the sophisticated and the best available AI solutions would have relied on the data available and advised against Sully's decision to land in the river," said a computer scientist advising companies on leveraging AI. He requested anonymity because one of the Indian firms is using his services currently.
"A true AI solution would have acted like Sully, using human intelligence and intuition," he added. The industry's latest obsession with AI needs an urgent reality check.
Perhaps, it may not be a bad idea for these companies to watch Stanley Kubrick's famous sci-fi movie called "2001- A Space Odyssey", which deals with AI and the humans vs robots debate.
There, an artificially intelligent robot called HAL 9000 takes matters into its own hands and its intelligence evolves beyond what the astronauts on the space mission had anticipated " so much so that HAL 9000 decides that its own survival at the cost of the lives of the astronauts would be crucial for the success of the mission. This is illustrated best in the film's famous "lip-reading" scene where HAL 9000 discovers that the astronauts are planning to deactivate it by simply reading the movement of their lips during a conversation in a separate room behind closed doors " despite never having been programmed to lip-read.
http://timesofindia.indiatimes.com/...ial-less-intelligent/articleshow/48051278.cms
Pankaj Mishra,ET Bureau | Jul 13, 2015, 01.02 PM IST
Over past few months, TCS, Infosys and Wipro have been pitching their versions of AI solutions as the next big thing in tech.
Indian IT's AI: More artificial, less intelligent
Over past few months, TCS, Infosys and Wipro have been pitching their versions of AI solutions as the next big thing in tech.
Depending who you ask in India's over $146-billion software industry (and that includes top executives at the biggest companies) the answer can be anything from automating repetitive tasks for clients, to applying predictive analysis for ensuring computer infrastructure doesn't crash down and so on.
Last month, India's largest software company Tata Consultancy Services launched Ignio, its new AI platform. The company talked about how the algorithms that form "the neural brain" of Ignio is capable of performing repetitive tasks. It can also predict peak demand for computer servers and ration storage capacity accordingly, in advance.
Wipro, the country's third largest software company is touting its cognitive computing system, Holmes, as "the Watson killer", referring to IBM's Jeopardy-winning supercomputer that's now being used by clients as an AI solution.
Executives at Infosys, India's second largest software firm, too are using the term AI while describing their current software platforms, most of which are focused on solving problems that occur repetitively.
If John McCarthy, or an Alan Turing were to hear these definitions on what AI can do, they wouldn't be too pleased. McCarthy, a Stanford scientist who originally coined the term "artificial intelligence" way back in 1955, defined it as "an artefact that could actually replace human level intelligence", clearly distancing it from machine learning.
A true AI solution has to be the closest to how a human brain thinks. And it will need to pass "the Turing test" wherein a machine exhibits the same intelligent behaviour as a human when faced with a problem.
As a top scientist advising companies on leveraging AI tools described, an artificial intelligence solution should be able to tackle problems not encountered before.
Over past few months, TCS, Infosys and Wipro have been pitching their versions of AI solutions as the next big thing in tech. These companies are using terms like "neural networks", "mimicking the human brain" and "applying human-like intelligence" in solving the tech world's biggest problems.
Most of the AI applications being touted by India's top software companies are based on using insights from accumulated data to take decisions. These solutions sift through massive chunks of data, identify repetitive patterns, and ensure that the systems automate these tasks.
This is hardly AI.
Let's take the example of the famous Hudson river plane crash in January 2009 when its pilot Sully Sullenberger landed the plane in the water. If he had listened to air traffic controllers, or relied completely on the available data for deciding whether to return to LaGuardia airport or land in New Jersey, he would never have landed the plane in the Hudson river.
As it turned out later, his decision to not rely on automated systems and instead go by his own judgement saved 155 lives.
"Most of the sophisticated and the best available AI solutions would have relied on the data available and advised against Sully's decision to land in the river," said a computer scientist advising companies on leveraging AI. He requested anonymity because one of the Indian firms is using his services currently.
"A true AI solution would have acted like Sully, using human intelligence and intuition," he added. The industry's latest obsession with AI needs an urgent reality check.
Perhaps, it may not be a bad idea for these companies to watch Stanley Kubrick's famous sci-fi movie called "2001- A Space Odyssey", which deals with AI and the humans vs robots debate.
There, an artificially intelligent robot called HAL 9000 takes matters into its own hands and its intelligence evolves beyond what the astronauts on the space mission had anticipated " so much so that HAL 9000 decides that its own survival at the cost of the lives of the astronauts would be crucial for the success of the mission. This is illustrated best in the film's famous "lip-reading" scene where HAL 9000 discovers that the astronauts are planning to deactivate it by simply reading the movement of their lips during a conversation in a separate room behind closed doors " despite never having been programmed to lip-read.
http://timesofindia.indiatimes.com/...ial-less-intelligent/articleshow/48051278.cms
Last edited by a moderator:
