The Slippery Slopes of the Digital Highway

Facts, Unchecked

AI (Artificial Intelligence) is dangerous. By that I don’t mean the familiar concern of machines taking over humanity. I refer to the now, when humans create and control technology. AI - like ChatGpt and its siblings - has an insidious effect on people.

We absorb things as truth often because of how they are presented to us, and because our hungry (and lazy) minds want quick answers - answers of the type that once we have them, we feel comfortable not to wonder about the question again. We want Truth so we don’t have to think again.

We absorb the printed word, marketing, Wikipedia and even religious texts, as facts, granting it permission and status as truth. We ingest summaries, headlines, essays, studies and curated stories, thinking we understand and are ingesting facts.

Wikipedia condenses decades of scholarship into one tidy article. But the scholarship can be from a disturbed person in a dirty basement, or from an Oxford professor. The Bible condenses centuries of stories into sacred narrative - curated, filtered, and produced by humans exercising editorial choice.

What makes AI worse though, is that ChatGpt, with its invisible herculean hands, grabs this long-standing habit of accepting information as truth simply because it is presented confidently, conveniently and with authority, and lifts it to the unquestionable heavens, from where articulate authoritative text seemingly anchored in existing factual knowledge is instantly produced, while simultaneously tailoring it to the conversant.

Accountability to unbiased verification is not built into the experience.

Side note about AI: It’s not artificial. The back end of whatever communicates with you, is an engine that pulls from very carefully curated-by-humans information. It has been programmed what constitutes a good answer. The “AI” machine is released when its programmers have taught and tested the engine enough times to the extent they are comfortable with its ability to answer the way they want it to. ChatGpt condenses thousands of pages in seconds - selected pages, words and patterns based on a human’s algorithm or bias, written sometimes by unqualified people.

As to how we absorb this AI: Most people accept what’s presented as truth and rarely stop, question, and most importantly, verify.

Who’s got the time? I want to and can get an answer in two seconds. I should take two weeks to study and prove an answer is true? Someone else did that work already (or so I am inclined to decide). If AI or Wikipedia or the bible says something, enough people have enough times accepted that, so anything it says is good enough as authority for me. (Not).

This isn’t naïveté — it’s human nature. Attention is easily grabbed; verification is work. Marketing exploits that. Politics exploits that. Social movements exploit that. Religion exploits that.

Millions anchor their lives, communities, and morals, on texts that are curated collections of who knows what - not double-blind verifiable and duplicable truths. Yet the words function as facts — because people accept them, uncritically, group by group, generation after generation.

The danger is not necessarily falsehood. It’s the process - accepting things as fact without scrutiny, letting unverified and perhaps unverifiable narratives shape lives, decisions, and societies.

A “study” “showed” this. A “survey” “showed” that. “Most people” do this.

Phrases often tease as universal fact. Distorted, filtered, slanted, partial, persuasive half-truths … are everywhere, presented as fact.

AI, Wikipedia, marketing, and religion all ride the same vector: easily digestible, authoritative, unquestioned.

“Well, what am I supposed to do?” asks the attention-deficit-disordered gentleperson. “Do my own research, look up information sources, check them out, go to a lab, become a scientist myself…?”

It depends on what you plan to do with the information you just read. And regardless, it’s at least important to understand that the text’s accuracy is highly questionable. Take it with a pound of salt. Most don’t. And that’s AI’s insidious danger: exploiting shortcutting, bypassing scrutiny, and fabricating “fact”, faster, deeper, wider and smoother than has ever been done before.

We engage mostly epidermally. Many people may have stopped reading this text many words ago, their attention fleeting. Most take soundbites as the entire meal and trust it’s healthy food.

You, though, who are still reading this, have done what many don’t do – carefully read, to fully understand and analyze. Whether you agree with me or not, this piece was written for you, the rare, attentive reader.

Hello there.