OpenAI Sunsets Generative AI Text Detection Tool

Open AI
(Image credit: OpenAI)

ChatGPT developer OpenAI has quietly shuttered its "AI Classifier" tool, which it introduced to the world mere months before. As AI Classifier was developed with the sole purpose of detecting generative AI content outputted by its own software, its shuttering is significant in the context of the multiple copyright lawsuits OpenAI is already juggling. The reason cited for the shuttering of AI Identifier rests with its low accuracy, which hovered at a paltry 26%. The move comes mere days after a number of AI-invested companies (including OpenAI) vowed to do just the opposite.

The main issue here stems from AI's opaqueness; an operational black box that seemingly means that being an AI product developer isn't enough to allow you to identify your product's output (which is as basic a capability as it gets). 

While there's been a number of advancements in "invisibly" watermarking the synthetic, generative output from AI-based image generators, the issue is a much more complex one when it comes to text. This happens because within actual text, there's no space to add any sort of identifying data; while an image allows for a number of different elements to be invisibly present (to our human eyes), any changes in the image, even invisible, add data that can then be picked up by specialist tools. 

But text is a wholly different beast, and there's also the issue that for someone truly devoted, they could simply re-type the entirety of a Large Language Model's output to keep only the text information, making fun of any invisible watermarks that might also be copied with a simple "copy and paste" command.

This is part of the reason why most "AI text" detector apps — such as the now buried AI Identifier and even GPTZero — process the actual words within a text, identifying writing patterns. Then it's really all about linguistics. If there's one thing that AI still can't do, it's to generate true "novelty." Everything that comes is a remix and a rematch of what came before (which justifies the usage of the word "synthetic" to describe AI-generated content, or current AI being described as a "stochastic parrot").

Analysing and identifying AI-generated text, then, seems to build upon the information theory work pioneered by Claude Shannon, which described the essential valuation of information as happening in the field of "surprise": if the information is surprising, that is, if it adds to or subverts an expectation, then it's relevant information (being more or less relevant according to how more or less surprising it is).

Just think about these two different titles for a Tom's Hardware article for the next generation's best GPUs: "Nvidia keeps the gaming performance crown in its latest generation" and "AMD steals gaming performance crown from Nvidia" (let's not get into whether those are good or badly written titles). The point is, one of those would be more surprising than the other and so the information's impact would also be higher.

AI detection tools such as the now defunct AI Classifier tool have essentially taken Claude Shannon's work and applied it to linguistics; because language follows patterns (some words work in some places and some contexts, but not in others), these AI tools look for patterns that have a low "surprise" value, that is, that don't deviate from the "baseline" amalgam of language that the model has arrived at after its training and weighting are complete (which is more aligned with the mean content of the model rather than the rest of its data).

Essentially, in the current world of AI "writing" there are generally no wordy-word, exquisite nor seldom-seen word sequences or turns of phrase such as this (nor this writer's own tendency to write run-in sentences with an ever-higher number of parentheticals).

This lack of surprise in AI's "writing" patterns is what these tools attempt to detect; something they're currently not very good at, as proven by AI Classifier's abhorrent 26% hit rate. Even GPTZero is only slightly better at detecting AI output than a monkey randomly attempting to do the same job would be.

The shuttering of AI Classifier shouldn't be seen as an admission of defeat on OpenAI's part, but rather a return to the drawing board. It's clear that the tool was broken and insufficient, and inspired little confidence. There's damage to be done when content written by an AI isn't properly detected, yes; but there's arguably an even higher impact to an actual human's writing being wrongly classified as AI-written. Copyrights, sensitive or classified information, corporate secrets, deepfakes, plagiarism: all of these could be brought to the table.

There's a responsibility to watermark AI correctly, but there's an even greater responsibility to not get it wrong. That said, OpenAI's decision to remove the tool sounds like the best one. 

TOPICS
Francisco Pires
Freelance News Writer

Francisco Pires is a freelance news writer for Tom's Hardware with a soft side for quantum computing.

Read more
The OpenAI logo is being displayed on a smartphone, with the Microsoft logo visible on the screen in the background, in this photo illustration taken in Brussels, Belgium
Microsoft and OpenAI investigate whether DeepSeek illicitly obtained data from ChatGPT
Sam Altman at OpenAI Dev Day.
OpenAI has run out of GPUs, says Sam Altman — GPT-4.5 rollout delayed due to lack of processing power
Nvidia Grace Hopper superchips
Chinese AI company says breakthroughs enabled creating a leading-edge AI model with 11X less compute — DeepSeek's optimizations could highlight limits of US sanctions
Nvidia RTX AI PCs and Generative AI
Nvidia RTX AI PCs and generative AI for games — how the Blackwell GPUs and RTX 50-series aim to change the way we work and play
A demo of Nvidia LATTE3D, a model that converts text prompts into 3D models. In this case, "Cinnamon spice latte with whipped cream".
Nvidia celebrates its 2024 AI research milestones — text prompt 3D modeling, advanced motion reconstruction, and more
Google Gemini Advanced
Google's Gemini 2.0 Flash is reportedly capable of removing watermarks — Also found to generate AI celebrity photos
Latest in Artificial Intelligence
ChatGPT Security
Some ChatGPT users are addicted and will suffer withdrawal symptoms if cut off, say researchers
Ant Group headquarters
Ant Group reportedly reduces AI costs 20% with Chinese chips
Nvidia
U.S. asks Malaysia to 'monitor every shipment' to close the flow of restricted GPUs to China
Ryzen AI
AMD launches Gaia open source project for running LLMs locally on any PC
Intel CEO at Davos
At Nvidia's GTC event, Pat Gelsinger reiterated that Jensen 'got lucky with AI,' Intel missed the boat with Larrabee
Nvidia
Nvidia unveils DGX Station workstation PCs with GB300 Blackwell Ultra inside
Latest in News
RX 9070 XT Sapphire
Lisa Su says Radeon RX 9070-series GPU sales are 10X higher than its predecessors — for the first week of availability
RTX 5070, RX 9070 XT, Arc B580
Real-world GPU prices cost up to twice the MSRP — a look at current FPS per dollar values
Zotac Gaming GeForce RTX 5090 AMP Extreme Infinity
Zotac raises RTX 5090 prices by 20% and seemingly eliminates MSRP models
ASRock fixes AM5 motherboard by cleaning it
ASRock claims to fix 'burned out' AM5 motherboard by cleaning the socket
ChatGPT Security
Some ChatGPT users are addicted and will suffer withdrawal symptoms if cut off, say researchers
project-g-assist-nvidia-geforce-rtx-ogimage
Nvidia releases public G-Assist in latest App to provide in-game AI assistance — also introduces DLSS custom scaling factors
  • bit_user
    If there's one thing that AI still can't do, it's to generate true "novelty" - everything that comes is a remix and a rematch of what came before
    I'm convinced this isn't as difficult as it probably seems. Once a language model can separate content, syntax, and style, then it just needs to master the trick of novelty in one or more aspects. And what's novelty except a random idea which actually makes sense? So, it seems all you really need is the ability for it to identify styles, syntax, etc. that it deems as unfamiliar, as well as a some kind of "fitness" test, to reject combinations that are inconsistent or too "far out".

    I think image generators are more likely to be the focus of novelty than language models, as users demand content that's not likely to provoke copyright infringement claims, but who knows?
    Reply