Should we Fear or Embrace Artificial Intelligence?

Elon Musk, the mind behind Tesla Motors, CEO of Space X and a co-founder of Pay Pal generated lots of news starting last August from his comments on Artificial Intelligence, including a comparison to nuclear weapons. 

 

In October the billionaire tech guru warned an MIT audience, "With artificial intelligence we are summoning the demon" He went on to say, "In all those stories where there's the guy with the pentagram and the holy water, it's like yeah he's sure he can control the demon. Didn't work out…"

Adario Strange from Mashable wrote a sensible follow-up article to Musk’s comments that takes into account the perspectives of other leaders in the AI field, and at a recent BlueWater Labs NYC meet-up he declared: “Musk may have ready too much science fiction…”

AI has not escaped the attention of many tech giants with deep pockets. Google paid $ 400 million last January for a British start-up Deep Mind. Indeed, its stated mission is simple if not ambitious: Solve Intelligence. The new company combines the best techniques from machine learning and systems neuroscience to build powerful, general-purpose learning algorithms.

The MIT Technology Review reported that Deep Mind has unveiled a prototype computer that mimics some of the properties of the human brain’s short-term working memory. This apparently solves one of the great challenges of neuroscience to replicate the same kind of memory in silico.


The Deep Mind computer is a type of “neural network” that has been adapted to work with an external memory. The result is a computer that learns as it stores memories which can later be retrieved to perform logical tasks beyond those it has been trained to do. Wow.

The future of the mundane work commute?

Yet there is still confusion about the basic definition of AI. Some think of AI as a machine that learns a specific algorithm, while others talk about autonomous robots or self-driving cars.

The AI discussion has in fact created a whole new vocabulary unto itself. Last month Vanity Fair took on the AI debate in their article entitled “Enthusiasts and Skeptics DebateArtificial Intelligence.” Topics included the evolution from “soft” to “hard” AI, and the debate around “Singularity” - defined as, "A technological singularity is a predicted point in the development of a civilization at which technological progress accelerates beyond the ability of present-day humans to fully comprehend or predict."

An argument has come up between the "Singularitarians" versus their skeptics about the moment when machine intelligence will surpass the human kind. This fascinating, if not somewhat ethereal debate is raging among the "digerati" - the elite of the computer industry and online communities.


The Maginot Line for this inflection point between the opposing camps is simple: A computer will be able to pass for human by 2030. The test for this is known as the “Turing Test” in which an average human interrogator will not have more than a 70% chance of distinguishing a computer from a human after five minutes of questioning. A partial threshold was passed last summer at the Royal Society in London when a computer fooled 10 of 30 judges, or 33%.

Hans Moravec and robots.

But let’s take a moment for a reality check. One of the major stumbling blocks for AI research is called “Moravec’s Paradox,” which says things that are easy for people to do are extremely difficult for computers to do. 

As Moravec observed: "It is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility."

This problem does not appear to be deterring the likes of Apple (with Siri) or IBM’s billion-dollar investment in Watson – its cognitive computing platform that uses natural language processing and analytics. Watson processes information akin to how people think, representing a major shift in an organization’s ability to quickly analyze, understand and respond to Big Data. Watson’s ability to answer complex questions posed in natural language with speed, accuracy and confidence is transforming decision-making across a variety of industries.

Frankly I like the pragmatic approach to employing AI that one well-known company is taking: Amazon is currently unleashing a robot army to speed holiday package sorting in their million-square foot order fulfillment center in Tracy, CA. 

"Whether it's consumables or toys or electronics, with 3,500,000 items plus in this building, the odds are, pretty much anything you wanted was likely here," says Dave Clark, Amazon's Senior Vice President of worldwide operations and customer service.

Kiva robots at work at Amazon facility in Tracy, CA.

 At most warehouses, goods are stored on shelves, and it's up to humans to out stock or retrieve stuff. But with the technology that Amazon acquired when it purchased Kiva Systems in March 2012, the goods come to the humans. Orange robots the shape and size of ottomans zip to the shelves, lift up the desired goods and whisk them to stations where workers complete the packing process.

With this system, not only is there no need for warehouse workers to march for miles up and down the aisles collecting orders, there is no need for aisles at all. This means Amazon can squeeze 50% more product into its already massive warehouse.

I am not quite ready to sleep with lights on in fear of Skynet - the self aware evil intelligence system featured in the Terminator franchise. It served as the series main antagonist and succeeded in scaring the daylights out of me starting back in the mid 1980s. 

Wonder about their IPO valuation?

But I do plan to give the whole matter of AI further investigation, all the while hoping that Amazon promptly processes my last minute holiday gift purchases this year, as I will surely wait until December 23rd to place my orders.