AI Text Classifier by OpenAI

External

OpenAI's AI Text Classifier was a free tool that predicted the likelihood of text being generated by AI models like GPT, aiming to promote AI literacy and combat misuse such as misinformation or academic dishonesty. It offered a simple interface for classifying English text with confidence scores, though it was never fully reliable. Discontinued in July 2023 due to low accuracy, it's now primarily of historical interest for understanding early AI detection efforts.

Pricing
Starting at USD6/moView pricing
CategoryAI Detection & Anti-Detection
AI Text Classifier by OpenAI

Description

OpenAI's AI Text Classifier was a free tool that predicted the likelihood of text being generated by AI models like GPT, aiming to promote AI literacy and combat misuse such as misinformation or academic dishonesty. It offered a simple interface for classifying English text with confidence scores, though it was never fully reliable. Discontinued in July 2023 due to low accuracy, it's now primarily of historical interest for understanding early AI detection efforts.

Key capabilities

  • Distinguish AI-written text from human-written text
  • Provide probability scores for AI generation

Core use cases

  1. 1.Verifying authenticity of student submissions
  2. 2.Detecting potential AI use in journalism
  3. 3.Promoting awareness of AI-generated content

Is AI Text Classifier by OpenAI Right for You?

Best for

  • Educators as a supplementary check
  • Researchers studying AI detection limits
  • Historical reference for AI literacy

Not ideal for

  • Primary decision-making on text authenticity
  • Short or non-English content
  • Current production use since shutdown

Standout features

  • Free public access (while available)
  • Fine-tuned on diverse AI and human text datasets
  • Adjustable confidence thresholds

Pricing

Prototypr Pro

USD 72/year

Prototypr Pro Lifetime

USD 250/one-time

Prototypr Pro

USD 6/month

User Feedback Highlights

Most Praised

  • Helped inform mitigations against AI misuse
  • Improved over prior classifiers on newer AI systems
  • Useful for longer English prose

Common Complaints

  • Low accuracy: 26% true positives, 9% false positives
  • Unreliable on short texts, non-English, or code
  • Discontinued due to poor performance
  • Easily evadable by editing