I Can Spot AI Writing Instantly — Here’s How You Can Too

About this video

• Video Title: I Can Spot AI Writing Instantly — Here’s How You Can Too
• Channel: Evan Edinger
• Speakers: Evan Edinger
• Duration: 00:15:59

Overview

Evan Edinger provides a comprehensive guide on how to identify text written by AI, offering practical tips and "red flags" to look out for. He explains that this skill is becoming increasingly important and shares his methods for distinguishing AI-generated content from human writing, using examples from various online platforms.

Key takeaways

• The M-Dash Misuse: AI models frequently use M-dashes (—) more than humans, making their overuse a subtle indicator of AI generation.
• Parallel Sentence Structures: AI often employs repetitive sentence structures, such as the "It's not just X, it's Y" format, which can be a tell-tale sign.
• The "Rule of Three": AI writers tend to structure lists or points in groups of three, a pattern that humans also favor but is often overused by AI.
• Uncanny Valley of Language: AI can sometimes use words or phrases in ways that feel unnatural or "off," creating a sense of an "uncanny valley" in the text.
• Generic and Vague Praise: AI-generated content often includes overly positive and generic compliments that lack genuine specificity.
• Unusual Analogies and Similes: AI may use analogies that are grammatically correct but feel forced or lack a natural connection.
• Repetitive Clarifications: AI can tend to restate points multiple times or provide unnecessary preamble, which can slow down the text and feel redundant.
• Human Writing Traits: Genuine human writing often includes personal anecdotes, tangents, imperfect grammar, and a unique "vibe" that AI struggles to replicate.