If you think about it, human intelligence is a fascinating topic. For example, depending on the source consulted, there are eight to ten different ways of classifying one’s intelligence quotient. That is why some people can solve complex formulas in their heads but have difficulty making small talk at social events. There is no mathematical formula to help them answer the question, “Do you like my new hair color?”
Yes, a high IQ does not always translate into a highly functional individual. Which brings me to the point of this piece, and the image on the left.
The graphic is a screenshot from my Facebook page. My post was, as I hope you can see, a tongue in cheek comment on the panic buying of toilet paper taking place around the United States. If you did not immediately pick up on the satire in the post, your IQ might be too high for your own good, but don’t abandon this piece yet. There is another point I want to make.
Again, if you have not already taken a good look at it, read the comment below the package of Angel Soft. That was Facebook’s response to my post.
Facebook, at least the version of artificial intelligence Facebook is using, was unable to recognize the sarcastic, hopefully humorous, message I intended. Instead, Facebook offered to help me engage in price gouging, which would likely be a crime in the middle of a pandemic.
The scary aspect of this situation is the reality we face in the not too distant future. Self-driving vehicles operating under the control of an AI have killed at least eight humans according to one source. Developers believe this issue can be overcome with time. Still, one must wonder how much damage will be done by automobiles and other devices operated by artificial intelligence before the systems are perfected.
Also, consider the comments made earlier about the highly intelligent. Sometimes, they can be as dumb as a stump outside their field of expertise. How will programmers with questionable interpersonal or social skills develop algorithms to recognize sarcasm, humor, idle threats, exaggeration, and other human idiosyncrasies?
Today, unless one disconnects all eavesdropping capable apps, limited forms of artificial intelligence monitor much of our communications. That is why mentioning a product in a social media post results in a deluge of advertising concerning products of that nature. Even having a casual conversation within the vicinity of a voice-activated device may result in a similar situation.
Given the increasing prevalence of devices equipped with some form of artificial intelligence, our dependence on them will increase. As that happens, given the problems already encountered, more and more opportunities will arise for an algorithm to miscalculate the appropriate response to something it encounters.
One can easily envision the day, if it has not already happened, when someone’s home or business is the focus of an embarrassing or troubling incident due to an intelligent app misunderstanding what it sensed. As the big brother aspect of government increases, imagine what would happen if an intelligent app misunderstood a dramatic outburst during a quarrel. Could the app, based on its programming, notify the authorities? If it does, will the 9-1-1 operator receiving the call realize it is a real-world version of Samantha in the movie “Her,” instead of a human?
Even more likely, is the expansion of incidents, with unacceptable consequences, such as the ones inspiring this piece. One event, of course, was the post mentioned above. Another was the unexplained blocking of a friend’s daily Bible blog. After sharing his blog for years, Facebook’s algorithms decided his blog was spam. Luckily, his complaints and inquiries reached a human being, and the matter was resolved.
The question now is, how long will it be before the computers get so smart they won’t listen to their human programmers?1 Maybe “The Terminator” was not so far fetched after all.
1Okay, that was a bit of hyperbole or sarcasm. Still, the idea that we can always control the devices we build is a bit of arrogance that might cause a lot of trouble in years to come. As anyone who has been the victim of a hacker, directly or indirectly knows, there are people out there trying to find ways to take over systems, or unleash havoc within them. Theoretically, at least, it should be possible to alter an artificial intelligence system to do whatever it calculates as appropriate.
© oneoldcop.com – 2020