ai-text-detector-github

To detect AI-generated text, you provide a passage such as an essay or email for heuristic analysis. The skill evaluates writing signals and outputs a risk estimate, confidence level, and caveats.

A text risk assessment evaluates whether a passage exhibits AI-like writing signals using heuristic methods. It returns a risk estimate, a confidence level, and caveats to help you determine if the text was potentially generated by AI.

Yes, analyzing text for AI-like signals requires a local Python environment with the ai-detector-skill library installed. The skill depends on this local library to execute its heuristic text analysis and generate risk estimates.

Text risk analysis provides an explainable risk estimate by identifying AI-like writing signals heuristically. Along with the risk score, it outputs caveats and suggested next steps to help you understand the assessment's limitations and context.

Heuristic AI text detection limitations include inherent uncertainty, which is why outputs include confidence levels and caveats. Since it relies on writing signal heuristics rather than definitive proof, it provides an estimated risk rather than absolute certainty.