Blog Issue #10, October 2025
Why we care about spotting AI-generated texts
People are naturally concerned about AI tells to spot AI-generated text because it is unnerving that AI can produce work that superficially resembles human writing.
Most people want a clear marker—something obvious to distinguish human vs AI writing—but this is not as clear-cut as we would like.
Image: Catharina Steel, drawing a round house layout similar to one she remembers drawing as a nine-year-old, October 2025
This trepidation is deeply psychological. Humans seek certainty and a sense of control, especially when faced with tools that could disrupt personal or professional domains.
Looking for surface AI tells like em-dashes, emojis, or stylistic quirks is often an emotional coping mechanism—a way to feel safe and competent in navigating new technology.
AI mimics human writing on many elements, including tone, punctuation quirks, and stylistic choices like em-dashes or emojis. Because of this, relying on surface AI tells is misleading; the real difference lies deeper, in context, nuance, and emotional depth.
There are two psychological reactions people tend to have toward AI:
- Fear-driven vigilance: People are desperate for AI tells, worried that AI will replace human creativity or work.
- Over-trusting: People who assume AI is inherently reliable or smarter than it really is.
In both cases, the response is rooted in emotion—fear or overconfidence—rather than understanding how AI works, its limitations, and the source or accuracy of its information.
The fear in creative industries
Writers, artists, and other creatives feel especially vulnerable because their work is personal, and the stakes are tangible: potential loss of jobs, opportunities, or recognition.
This fear drives the search for markers of human v AI writing, even though punctuation, emojis, or stylistic quirks are unreliable indicators.
The more productive approach is to understand AI’s limitations and the irreplaceable value of human insight in writing—the depth and judgment.
AI tells are an emotional reassurance—not a reliable indiactor
Looking for clear markers, from a psychological perspective, is about seeking reassurance—if you can spot AI, you feel safer.
However, AI mimics human writing—so it can include or omit these features depending on prompts. This has resulted in some writers avoiding perceived AI tells to prevent being accused of using AI-generated text.
Surface features are therefore emotion-driven clues, not reliable evidence.
Detection tools and AI tells
AI tends to produce surface-level writing and is often superfluous and generic. Text is grammatically correct but often lacks depth, subtlety, or contextual nuance unless a user carefully guides and/or edits it.
AI mimics human writing and emotion—but it doesn’t truly understand it. Cultural references, personal experiences, or layered meaning require a human to heavily prompt and direct for AI to include this human insight in writing in a way that contains depth and perspective.
Detection tools exist, but their reliability is limited:
- False positives: human-written text flagged as AI.
- False negatives: AI-generated text passes as human.
A concise or simple section written by a human may resemble AI output. Context, nuance, voice, and perspective are essential—surface appearance alone is not proof, so detection tools should guide, not dictate.
Dangers of over-relying on AI
Some rely on AI to fully research or draft work without verification. Psychologically, this signals overconfidence in the tool and a misunderstanding of its limitations.
To use AI effectively:
- Guide it actively.
- Edit heavily to ensure the correct perspective, context, and nuance.
- Verify the accuracy and relevance of all information AI provides.
AI saves time, particularly in research, but human oversight remains essential to ensure content is factual, contextually rich, and emotionally resonant.
Coming Up Next
Next month, I will be delving into the use of AI as a tool.
To read my previous post about Werewolves, Shapeshifters, and Weredogs, click here.
To read my author newsletters, go to: Substack