Six months after its launch, Open AI’s artificial intelligence detector tool “AI Classifier,” abruptly ended its service seemingly overnight.
OpenAI stated that their very own AI detector will no longer be accessible due to low accuracy rates.
The AI Classifier was previously intended to distinguish human writing AI-generated through analysis of linguistic features, sentiment analysis, and pattern recognition. However, things took a different turn due to the tool not being able to do so at a precise level.
“It should not be used as a primary decision-making tool, but instead as a complement to other methods of determining the source of a piece of text,” OpenAI asserted.
OpenAI openly claimed that their classifier is ‘not fully reliable’ from the first day it was launched, and upon their evaluation, it only identified AI-written text 26% correctly while incorrectly classifying human-written text as AI-written 9% of the time – and many writers in the digital community have also openly criticized this inaccuracy.
In addition to that, they also pointed out that the classifier is ‘very’ unreliable on texts below 1,000 characters, another reason they got rid of the tool.
It's quite comforting to know that OpenAI acknowledged their very own shortcomings and openly (albeit silently) let the world know they no longer support the classifier.
OpenAI did note they are currently developing a new and improved AI detection tool that would make up for the loss in their previously launched detector.
“Our work on the detection of AI-generated text will continue, and we hope to share improved methods in the future,” OpenAI claimed.
There are many other AI detectors on the market, tools like Originality and GPTZero have been at the forefront of the generative-ai detection wave, and only time will tell if these tools can also weather the storm or if they'll eventually succumb to accepting how great (and borderline impossible) it is to accurately detect AI-generating writing.