The Stupidity of Computers by David Auerbach

Auerbach starts off by stating that computers are brainless high maintenance machines that require babysitting and a step-by-step tutorial to execute any task. While a computer can access a magnitude of data to satiate any user’s request, any logic or common sense is lacking especially with regards to human interaction. The breakdown occurs in the computer’s understanding of the user and its expected task. Communication issues typically stem from “ambiguity inherent in a sentence’s syntax and semantics.” (Auerbach) This adds to the pressure of having to be overly specific in details and situations, thereby leaving no other possible interpretation of human language. These inhibitions challenged programmers to refine a computer’s intelligence level by improving its linguistic capabilities.

The author then proceeds to explain the history of the search engine and its progress since the 1960s. From one system to the next, each had its own barriers. It is true that early search engines produced a plethora of results, but these were disorganized and oftentimes randomly selected links that may or may not be relative to the user’s original desired outcome. When Google came along, instead of conquering the semantic issue, a few researches decided to bypass the issue altogether by enabling computers to identify the most appropriate results by the pages to which they are linked. This set Google apart as a progressive search engine by far surpassing other search engines that had the ability to locate more relevant pages with analogous information. The problem of a computer’s understanding was not solved, but at least communication improved and keywords produced a higher chance of relevant content.

Computers generate a lot of information, but it is ultimately up to humans to organize the information in a logical, coherent way. When left up to the computer’s discretion, items such as books, articles, or web pages are oftentimes categorized incorrectly based on its own criteria. When it comes to shopping and major online companies like Amazon, again categories are preset by humans. Amazon stores information about their users and their previously purchased goods, which makes their search engine capable of predicting future inquiries. However, there are still fallacies within the system that must be facilitated and edited by humans. 

From beginning search engines to Google, search engines over time have become more and more advanced, specialized, and sophisticated. However, a computer will never completely understand people nor will it be able to take the place of human intelligence. No computer has ever been able to pass Alan Turing’s intelligence test of being able to convince an audience it is human. (See implications for updated information) Auerbach comes to the conclusion that because computers will not fully be able to fully join our world, we will have to eventually join their world. As we become more fully integrated with computers and dependent on their powers, we will find ways to conform our ways of thinking to their ways of thinking. Auerbach concludes that humans will acquire the limitations of computers, thereby “dumbing” themselves down to a level equal of the machine.

Implications & Examples

One obvious implication is Auerbach’s assumption of the limitations of the computer. If the computer has progressed exponentially has he explained in his article, how could he be so certain the computer will never be able to create ontological categories? With the virtual world being as unpredictable as it is, nothing should ever be assumed. As Auerbach took so much time to expand upon, search engines have become increasingly refined with their classification abilities at a very rapid pace. Moreover, a computer cannot possibly pick up sarcasm, political slants, irony, or other nuances unique to the human race, right? Not necessarily. An Israeli research team developed SASI, a Semi-supervised Algorithm for Sarcasm Identification, which can detect sarcastic comments online with 77 percent precision. (http://www.cs.huji.ac.il/~arir/10-sarcasmAmazonICWSM10.pdf) New algorithms surface all the time closing the gap more and more between artificial intelligence and the human brain. Pattern recognition abilities have heightened a computer’s perception of humans. According to a recent study, scientists were successful in developing a program that could detect whether a person was faking injury or in real pain. (http://www.businessinsider.com/r-if-you-want-to-fake-it-dont-do-it-around-this-computer-2014-21) Emotion detecting robots have been around for a few years now, and the software is only being further perfected. 

Oddly enough, while I was writing this blog, for the first time ever a computer passed the 65-year-old Turing Test mentioned previously. Five supercomputers entered the 2014 Turing Test, and the winner, Russian computer “Eugene Goostman”, is certainly a milestone in the world of artificial intelligence. Eugene was able to fool 33 percent of humans it was a 13-year-old boy. (The threshold for passing is 30 percent.) The consequences associated with this seeming victory are unknown at the time. The reality of a computer being able to trick someone it is human is a tool that could be used in conducting cybercrime or combatting cybercrime. The creators of Eugene plan to continually make the machine smarter.

Ultimately David Auerbach should refer to the stupidity of humanity rather than the stupidity of computers. After all, it is the human’s decision to step down and join the computer’s reality. I personally do plan on dumbing myself down to a computerized version of myself. Now we are entering a phase of human embarrassment where people are eager to live their lives digitally and define their lives based on social media. We still have (relative) control over our computer dependency; to give up that control is our dumbness—not the computer’s.

References:
http://www.cs.huji.ac.il/~arir/10-sarcasmAmazonICWSM10.pdf
http://www.businessinsider.com/r-if-you-want-to-fake-it-dont-do-it-around-this-computer-2014-21
http://www.washingtonpost.com/news/morning-mix/wp/2014/06/09/a-computer-just-passed-the-turing-test-in-landmark-trial/
http://www.nbcnews.com/tech/tech-news/turing-test-computer-program-convinces-judges-its-human-n125786