Ruben Hassid Identifies AI Writing Pitfalls and Offers Solutions

In a recent LinkedIn post, Ruben Hassid delves into the increasingly common issue of AI-generated content that lacks authenticity and human touch. Hassid, a proponent of leveraging AI tools effectively, outlines specific linguistic markers that often betray AI authorship and proposes a set of “skills” designed to refine AI outputs, ensuring they resonate more naturally with human readers.

Identifying the Hallmarks of AI Writing

Hassid begins by cataloging common phrases and word choices that, in his view, signal AI-generated text. He argues that these elements, while perhaps appearing sophisticated to the AI, sound unnatural and overused when encountered by human readers. Among the “tells” he lists are words like “delve,” “crucial,” and “tapestry,” which he suggests are rarely uttered in natural conversation.

“Delve.” You’ve never said it out loud. Once.

The author further points to formulaic introductory and concluding phrases, such as “Hope this helps” and “After careful consideration,” as giveaways. Hassid also criticizes the tendency of AI to oversimplify complex issues with phrases like “Most people…” and to rely on corporate jargon like “robust” and “seamless.” He believes these patterns make content predictable and unengaging.

Skills to Enhance AI-Generated Content

To counteract these perceived shortcomings, Hassid introduces eleven “skills” that he personally uses with AI models like Claude to improve written output. These skills are presented as commands or prompts designed to guide the AI’s writing process, moving it away from generic patterns towards more personalized and effective communication.

Leveraging AI for Authenticity

Hassid’s proposed skills focus on several key areas:

• Drafting and Structure: Skills like `/writer` aim to generate initial drafts that adhere to the user’s intended structure, rather than the AI’s default.
• Conciseness and Clarity: The `/editor` skill is designed to “cut the fluff and the filler,” ensuring every sentence has a purpose.
• Accuracy and Trust: Hassid emphasizes the importance of a `/fact-checker` to prevent the inclusion of “invented stats and fake-sounding quotes,” noting that “One hallucinated number and nobody trusts you again.”
• Style Emulation: Notably, Hassid includes a `/sound like your posts` skill, which requires feeding the AI examples of previous work to mimic the user’s unique voice, a feature he suggests can replace many paid “humanizer” tools.
• AI Pattern Avoidance: Specific skills like `/ban-the-AI-words` and `/ban-the-AI-patterns` are intended to automatically remove the tell-tale phrases and sentence structures that mark content as AI-generated.

Hassid argues that by implementing these targeted skills, users can achieve a significant reduction in “AI smell,” resulting in content that is both polished and authentic. He sees this as a crucial step for professionals aiming to maintain credibility in an era of increasingly sophisticated AI writing tools.

The patterns out you faster than the words do.

Ultimately, Hassid’s post serves as a practical guide for individuals looking to harness the power of AI for content creation without sacrificing genuine human connection. He provides a clear framework for identifying and rectifying common AI writing deficiencies, offering a pathway to more authentic and impactful digital communication.

12 tells deleted. 11 skills. 0 AI smell.