A recent post on Reddit noted a problem a job candidate had. He was highly favored as a hiring pick, but then the company's CEO decided the applicant's hiring essay was written with artificial intelligence (AI). To add insult to injury, the company was accusing the candidate of trying to apply for a job area he didn't apply for in the first place, sales. He was struggling trying to decide how to respond, whether to still take the company serious for a job prospect, or just drop them altogether.

The burning question on everyone's mind, however, was, what exactly did this CEO use to assume the candidate's essay with written with AI? Most thought he was an arrogant idiot who couldn't identify a good writer from a computer.

Factors That Tip Off AI Usage (Supposedly)

From news articles to blog posts, AI can create text that mimics human writing to a remarkable extent. However, there are ways to discern whether an article has been written by AI.

It's All About the Tone

One of the primary characteristics of AI-generated articles is, supposedly, a consistent tone and style throughout the piece. Unlike human writers who may exhibit variations in writing style and voice, AI models often adhere to a standardized pattern. They lack the nuanced touch that a human writer can bring to their work. So, if you notice a monotonous or rigid tone in an article, it might be a clue that it was produced by AI.

Um...this completely disregards the fact that writers for decades have been formally trained to have consistency in tone, depending on the audience of their document and the intent of the written product. It's bit like saying good cars don't have wheels, but every functional car people use rolls on wheels.

Lack of Personal Experience or Emotion

AI-generated articles typically lack personal experiences or emotional anecdotes that human writers tend to include. While AI models can be trained on vast amounts of data, they lack the first-hand experiences and emotional depth that human authors bring to their writing. If the article seems devoid of personal stories, subjective viewpoints, or emotional nuances, it may be indicative of AI-generated content. This is sort of half-true. If you've ever read corporate or government writing, it's about as generic and void of life as a rock on the moon.

Flawless Grammar and Syntax

AI models are trained on large corpora of text, enabling them to generate grammatically correct and syntactically sound sentences. Consequently, AI-written articles tend to have impeccable grammar and structure. They are less likely to contain common mistakes such as typos, punctuation errors, or awkward phrasing. If you come across an article that seems too perfect in its linguistic construction, it could be a sign that AI was involved in its creation. This is likely the mistake that corporate managers like our boy scout CEO mentioned above thinks is a clear sign of AI. Honestly, this could be taken one of two ways: 1) there is general assumption now that the general population is so bad at writing they are guaranteed to make mistakes left and right, regardless of using Microsoft spellchecker, Grammarly and whatever else is out there for good writing assistance in document editing, or 2) the level of writing is now so low, anyone who writes above a 4th grade level is assumed using a computer to do it and can't possibly be intelligent enough write competently.

I actually ran into this problem years ago in my first college classes, which leads me to think the perception of AI is both. It was my freshman year, but I had been a hardcore writer in high school. I submitted my first essay paper and professor immediately accused me of cheating and hiring someone to write the paper. So I challenged him back to give me a topic on the spot and 30 minutes to write it ad hoc. After ten minutes he gave up and accepted that I could actually put a sentence together properly.

Absence of Bias or Subjectivity

Human writers bring their perspectives, biases, and subjective interpretations into their work. AI, on the other hand, lacks personal opinions or biases unless intentionally programmed to simulate them. If an article seems devoid of subjective analysis or does not reflect a particular stance, it might indicate that it was generated by AI.

Again, this runs counter to good writer training. In fact, the best writers and analysts are able to write from any angle of an issue, without giving away how he or she really feels about the matter. It assumes humans are so stupid, they can only do one thing at a time. Granted, some folks might think of exactly that kind of a person they know immediately, but the general mid-range of humanity is capable of multi-tasking a bit.

Unusual or Incoherent Content

Although AI models are capable of generating coherent text, they can sometimes produce content that seems unusual, incoherent, or off-topic. These anomalies can occur due to errors in training data, model limitations, or contextual misunderstandings. If you encounter an article that seems to jump between ideas abruptly, contains nonsensical statements, or appears to lack logical coherence, it could be an indication of AI involvement.

So far, I've seen humans write more incoherent babble than computers. Just look at anything on Reddit for proof of that hypothesis. That said, I do agree that computers in translation of languages definitely give themselves away. There are specific ways that words are positioned that a native speaker/writer knows and a computer cannot emulate. This is why cheaters using online translators always get caught in their language classes. That said, using AI to clean up an existing translation for better interpretation has proven to be very good at getting closer to native speaker/writing status.

So, What Can We Conclude?

Human pattern is probably the most indicative evidence of AI use versus actual human product. AI doesn't follow a specific pattern. So one article on a topic can sound very professor-like, even archaic in word choice, and a second one a minute later could use very common language word choices instead. Human writers stick to their patterns and word choice, producing them again and again over time. No CEO or interviewing manager is going to see this with one writing sample, but it can be seen in a few days when a number of documents by a supposed author are examined together.

By paying attention to factors such as obvious robotic tone, repeat grammar choices, and unusual content, readers can develop a critical eye over time to discern AI-generated content from their reports or people they work with.