The video explains how to manually humanize AI-generated content by identifying and addressing six key indicators—such as surface-level writing, overused transitions, lack of nuance, the rule of three, burstiness, and overly complex language—to make the text more authentic and natural. Through a step-by-step editing example, the presenter shows how these techniques effectively bypass AI detection tools, resulting in content that appears genuinely human-written.
In this video, the presenter demonstrates how to manually humanize AI-generated content to bypass AI detection tools effectively. The process begins with understanding how AI generates content, which often results in surface-level writing that lacks depth, nuance, and originality. The presenter uses a metaphorical hat containing six key indicators to look for when reviewing AI content: surface-level understanding, overused transition words, lack of nuance, the rule of three, burstiness in sentence structure, and overly complex or wordy language. By identifying and addressing these elements, one can transform generic AI text into more authentic, human-like writing.
The first step involves recognizing surface-level content, which typically offers a basic, generic overview without detailed insight or specific data. The presenter emphasizes the importance of deepening this understanding by adding personal knowledge and context to each sentence. Next, the video highlights the overuse of certain transition words like “however,” “therefore,” and “in conclusion,” which AI tends to insert excessively at the beginning of sentences. Removing or replacing these with more natural phrasing helps make the text sound less mechanical and more conversational.
Adding nuance is another critical aspect, as AI often presents information in a definitive, matter-of-fact manner without acknowledging exceptions or complexities. The presenter suggests incorporating edge cases or qualifiers to reflect the true nature of academic writing. Additionally, the “rule of three”—a common AI pattern where three items are listed for stylistic effect—is identified as a giveaway. Breaking up these triplets or rephrasing them can reduce the AI-like feel of the text.
The video also discusses “burstiness,” which refers to the variation in sentence length and complexity. AI-generated paragraphs often have uniformly long or similarly structured sentences, lacking the natural rhythm found in human writing. Introducing sentence length variation, including shorter sentences, enhances readability and authenticity. Lastly, the presenter warns against overly sophisticated or wordy language that AI tends to use, which may not align with the expected writing style of a particular academic field or level.
To illustrate the process, the presenter walks through editing an AI-generated paragraph about organic photovoltaic (OPV) devices. By applying the six indicators, they add nuance, simplify language, remove repetitive transition words, break up sentences, and avoid the rule of three. This careful, sentence-by-sentence revision transforms the text into a more natural and credible piece of writing. The final step involves rechecking the revised content with AI detection software, which then confirms the text as likely original, demonstrating the effectiveness of manual humanization techniques.