study claims that AI-generated content can be detected.

Search Logistics, an SEO agency, has published a report claiming that almost 90% of CNET’s AI-generated content could be detected using a public AI detection instrument.

CNET Money was using an “AI assist” to create explainers to answer frequently asked questions. They had published 75 of these articles by mid-January.

Why do we care? Search Logistics reports results that are consistent across large amounts of text. This could help answer many of the questions about ChatGPT-like content creation and management tools. One thing is that Google has stated it will consider AI-generated content “spam” threatening search rankings for websites that heavily rely on such content. Copyright can only be granted to human-generated content, as has been the case with Google’s Copyright Office.

These positions beg the question: Is AI-generated content possible to identify reliably? According to the Search Logistics study, yes. However, this doesn’t mean AI cannot replace human content creators.

The data. This report showed that:

• CNET detected 87.2% AI-generated content.
• 12.8% were not detected.
• 19.2% of articles tested contained at least 50% content that was generated using AI.
• 7.7% of respondents had AI-generated content with 75%+.

CNET stated that AI-generated content was fact-checked by humans and edited by them.

The post AI-generated Content is Detectable, New Study Claims was first published on MarTech.