While automated AI detectors are not reliable, humans can sometimes detect AI writing based on changes in writing style or tell-tale signs in the text. However, as per AI analyst and Wharton professor Ethan Mollick, AI writing is currently undetectable and likely to remain so. He advises against the use of automated AI detection tools due to their high false positive rates.
Key takeaways:
- OpenAI has admitted that AI writing detectors do not reliably distinguish between AI-generated and human-generated content, often leading to false positives.
- OpenAI's AI models, including ChatGPT, do not have the ability to know whether text is AI-generated or not, and can often generate false or misleading information.
- Despite the unreliability of automated AI detectors, humans can sometimes detect AI writing, especially if they are familiar with the typical writing style of the supposed author.
- AI analyst and Wharton professor Ethan Mollick advises against using automated AI detection tools due to their high false positive rates.