search engine optimization company Search Logistics has launched a report that claims nearly 90% of CNET’s AI-generated content material was detectable utilizing a public AI detection software.
CNET Cash had been experimenting with an “AI assist” to compile explainers in response to regularly requested questions. By mid-January they’d revealed around 75 such articles.
Why we care. The outcomes reported by Search Logistics, if iterable throughout bigger samples of textual content, could possibly be related to the various questions which have been raised about using ChatGPT-like content material creation instruments. For one factor, Google has mentioned it’ll regard AI-generated content as “spam,” thus threatening search rankings for websites that come to rely closely on such content material. The Copyright Workplace has persistently mentioned that solely human-generated content material will be copyrighted.
Such postures beg the query: Can AI-generated content material be reliably recognized? The Search Logistics examine suggests the reply could also be sure. This doesn’t essentially imply AI can’t exchange human content material creators; simply because the AI detection software (Originality.AI) is aware of when it’s being fed the ruminations of a robotic, it doesn’t comply with {that a} human reader can inform.
The information. The report discovered that:
- 87.2% of CNET’s AI-generated content material was detectable.
- 12.8% averted detection.
- 19.2% of the articles examined had 50% or extra content material generated by AI.
- 7.7% had 75%+ AI-generated content material.
(CNET has mentioned that AI-generated content material is fact-checked and edited by people).
Get MarTech! Day by day. Free. In your inbox.