Sure, it Sounds Good, but Is It Accurate?

AI said this was English. I said, ‘Sure, buddy.’

I make a living being accurate.

My job is to present information in a non-fiction format that is both accurate and provocative.

Let’s examine some synonyms of provocative (adj.): exciting, charged, provoking, motivational, motivating, stimulating, edgy, inciting, motivative, inflammatory, explosive, inspiring, instigative, incendiary, inspirational, instigating, triggering, piquing, etc.*

I am tasked with finding the story inside the truth. My guiding principle? Facts.

The Challenge of Accuracy

This is not easy at times. Maybe the story is just not that interesting. Perhaps the story doesn’t have “enough legs” for the time slot allocated. Or possibly there aren’t enough facts to craft an honest story.

It’s this last point that sticks with me. In this modern age of the ease and convenience of answers at our fingertips, I see more and more that the answers are incorrect.

How frequently do you see a meme or a posting that catches your eye only to realize it’s completely made up? When’s the last time someone said to you, “Did you know…” and you thought that it seemed a bit off? Or when you look at an AI-generated image and it takes a moment to realize the lettering is gibberish?

This happens to me all the time and, in the completely unscientific way humans tend to behave, I assume I’m not alone. This assumption is called the False Consensus Bias.

“The false consensus effect is a psychological tendency to assume that our own opinions, values, and behaviors are more common than they actually are. This bias can lead us to misjudge the beliefs and attitudes of others, overestimate the agreement others have with our point of view, and make decisions based on flawed assumptions about group consensus.

This phenomenon stems from the human need to validate one’s own perspectives and the tendency to seek out and remember information that aligns with our existing beliefs.”

So how does this influence what we believe and what we don’t?

It can sometimes trigger that internal part inside of us that doesn’t want to hear the truth because we like what we’ve told ourselves.

And you know what makes this a lot easier? Social media and AI.

Where AI Trips Up

I regularly find mistakes in searches through ChatGPT, Google’s Gemini, and Meta’s AI mode. Have you ever seen the AI-generated summary on some Facebook posts? They’re hilarious.

The post says: “Imagine if John Wilkes Booth didn’t go to Ford’s Theater.” And the AI assistant says: “President Lincoln's reactions to the play.”**

It’s not just gibberish language and six-fingered men not named Count Rugen. It’s data generated from a wide variety of sources, but not always processed accurately.

It may sound good—but dear lord, it is not accurate.

Hell, look at how Google’s Gemini describes it:

“No, AI does not always give accurate information; while AI models can produce high-quality outputs, they can also generate misinformation, false information (hallucinations), or inaccurate predictions due to limitations in their training data, biases within the data, and the inherent nature of their probabilistic outputs. To ensure accuracy, it is essential to fact-check and verify AI-generated content with other reliable sources, as no AI system is 100% accurate.”

Wrapping It Up

It may sound good—those memes, posts, and AI search results—but don’t automatically accept it as accurate.

It may confirm the false consensus bias we all fall victim to.

But facts and the truth are vital, not only for documentary filmmaking but also for communication overall.

Notes

[*] For those wanting a reference, I used m-w.com and only selected the orange and peach words.

[**] I made this up, but I did see an equally ridiculous one about a college football game.