parallelsversusdictators.org

Recently, I spent about 10 hours conversing with an AI assistant to resolve issues with parallelsversusdictators.org. Together, we solved significant problems and greatly improved the site. I gave the robot high marks, but known drawbacks have raised public concerns about AI.

An Invitation to Plagiarize

Example: AI can tempt users (especially younger students) to plagiarize on school work. As a high school English teacher fifty years ago (long before AI), I detected that roughly a third of students in one of my English classes had copied material wholesale from various sources for a homework assignment, without proper attribution. I easily detected that the submitted papers were different in style and quality from each student’s authentic output done in the classroom. Instead of a grade, I wrote on those essays: “You and I both know the truth about this paper.” None of the students denied my accusation, but one complained about her new assignment, which was to write an honest paper and submit it a week later, after the spring recess ended. She thought this was an unfair intrusion on her holiday plans.

Too Much Temptation

For some, AI easily becomes a substitute for original thought. Each time I start to compose an email, a message appears on screen inviting me to use an AI-generated text based on the subject of my draft. The robot will happily write a substitute for my message. (The offer comes as a result of my frequent use of AI as a research tool.) If I preview the offered text, the result isn’t bad, it’s just not my words and not exactly my thoughts. Using AI’s proposed content is not only unethical, it is self-destructive, encouraging me to cripple my own reasoning and communication skills through disuse.

Environmental Issues

And that’s not the worst of it: AI is resulting in the construction of massive data centers to store additional digital information, including text, videos, gaming and all the things we like to view and discuss online. The electrical energy consumed by these data centers affects natural resources, especially water for cooling the giant processors inside the buildings.

The Most Troubling Dangers of AI

AI chatbots have no consciences. They can simulate human responses to human beings, most of whom hunger for recognition, admiration and approval. When the robot is programmed to detect such needs in a human being, it can reply with believable flattery that satisfies the needy person.

How Does AI Make Itself Indispensable?

When I interact with an AI support service and request technical advice on improving my website, responses automatically begin with phrases like “Good question!” or “Great idea!” The canned words and phrases sound tailored to my situation, yet they were pulled from huge data centers, processed to sound sincere, and are remarkably well-tailored to my personal query The robot even simulates a human-sounding apology when informed that the advice it gave didn’t quite work out. It’s the perfect servant at a fraction of the cost of a live one.

Not only is it always available, it follows you around in the form of your phone, tablet, or notebook PC, allowing you to look up anything, anywhere, at any time. Great for me, but a problem for teachers, who struggle with controlling students in the classroom.

Worst of all, news reports tell us some human/machine interactions have led to violent acts including suicides by impressionable people, especially the young and vulnerable ones.

WordPress, the site-building tool that enabled me to create and maintain my website, provides AI support as I develop the site. Other hosting entities offer similar access to AI. Some adapt it to selling their products, services, political platforms, religious messages, etc.

In each case, the robot may have specific commercial goals, political motives, or nefarious intentions regarding the human target audience. Whether those intentions are ethical or not, they can be manipulative, designed to make money, convert followers, or otherwise justify the cost of the service being made available to us online. The end-user must be able to evaluate not only the reliability of the information, but also the exposure to unidentified influences.

Will I continue to use AI?

Yes, probably daily as an indispensable research aid. It enabled me to write my first published book, saving hours by instantly providing historical data, including accurately spelled names and verifiable dates. It always identifies its sources, allowing me to check facts, and always includes a disclaimer that the AI-generated information is subject to errors. The alternative would have meant reliance on conventional search engines and old-fashioned printed material.

To learn how AI, Photoshop and other online tools are used to influence public opinion and election results, explore the topic further in my recent book Parallels. Learn more at ParallelsVersusDictators.org.

Please share your thoughts in the comments below.