AI Hesitates to Release 99.9% Accurate ChatGPT Detection Tool
Despite high demand from educators, OpenAI is concerned about user backlash and technical circumvention.
- The tool can accurately detect ChatGPT-generated text with near-perfect precision.
- OpenAI fears the tool could deter nearly 30% of its users from using ChatGPT.
- Technical workarounds, such as rewording or translation, can bypass the watermarking.
- The company is exploring alternative methods like metadata embedding for AI content identification.
- Concerns include potential stigmatization of non-native English speakers using AI tools.