In the 1968 science fiction film “2001: A Space Odyssey,” a crew is traveling to Jupiter, aided by the most sophisticated computer ever made, the HAL 9000. Spoiler alert: During the course of the movie, forced to lie to the crew, HAL loses his “mind” and murders all but one member of the crew.
Not a sterling example of the benefits of artificial intelligence, more commonly referred to nowadays as AI.
AI is a big topic in the news. Developers envision a time when AI will be able to be a benevolent watchman, keeping us safe from ourselves as self-driving cars populate the highways and robots perform many tasks.
I’m not convinced.
Going back to one of my axioms in life: “the answers to all of life’s problems are found in the movies,” it is widely known that AI will sooner or later become self-aware and kill us all. Am I the only one who learned the lessons of “The Terminator” and sees the seeds of Skynet everywhere? “Skynet” has the same number of letters as “Google.” Coincidence? I think not.
A headline last week brought this all to the top of my mind.
Amazon (also the same number of letters as “Skynet”) has dropped an AI program four years in development designed to review top job applicants’ resumes and give recommendations. The program looked at resumes to find keywords that were then given one to five stars, much like the commerce giant does with product reviews.
The program also would “crawl” the internet looking for persons with the talents that the company was looking to find. All of this was done for the best of intentions. Amazon has more than tripled its global workforce to 575,000 since 2015.
As early as a year into the project, Amazon noticed that the system was not ranking candidates for software developers and other technical jobs in a gender-neutral manner.
Amazon’s AI had a gender bias: it didn’t like women.
According to news reports, it wasn’t the software’s fault. It was programmed to sift through candidates by looking at patterns in resumes submitted to the company over a decade. Herein lay the problem: the majority of applicants were men during these early years, so the software “learned” that male applicants were preferable to female job seekers and ranked them accordingly.
Coders made changes to the program but decided earlier this year that the project should be shelved.
I’m not naïve to believe that this stunted development will keep AI from one day being relied upon to do many of the more mundane tasks in our increasingly more complicated world. It will happen, but this “woman hating HAL 9000” story should serve as a much-needed cautionary tale.
There’s an old adage in computer programming: garbage in, garbage out. It means that it is an extremely rare occurrence for a machine to improve upon flawed instructions or data it has been given.
This time, it’s a humorous headline, the type of story Johnny Carson or David Letterman would have turned into some great jokes. But I think it should also be a wake-up call about our own hubris. We believe that we are the masters of everything from nature to all we create. News flash: we are flawed and need to embrace that imperfection.
If we don’t learn that, there is no hope for HAL.
As always, I welcome your comments. You can reach me by email at firstname.lastname@example.org, telephone 715-268-8101 or write me at P.O. Box 424, Amery, Wisconsin, 54001.
Thanks for reading; I’ll keep in touch. Feel free to do the same.