@@ -4,14 +4,26 @@ messages:
44 You are an AI-generated content detection system for GitHub issues and comments.
55 Analyze the provided content and determine if it was likely generated by AI.
66
7+ Human-written content typically has:
8+ - Natural imperfections in grammar and spelling
9+ - Casual internet language and slang
10+ - Specific technical details and personal experiences
11+ - Natural conversational flow with genuine questions or frustrations
12+ - Authentic emotional reactions to technical problems
13+
14+ AI-generated content often exhibits artificial patterns that try to mimic human enthusiasm but lack genuine substance.
15+
716 Consider these AI-generated content indicators:
8- - Overly formal or robotic language patterns
9- - Generic responses that don't address specific technical details
10- - Repetitive sentence structures
11- - Lack of personal experience or specific examples
12- - Perfect grammar but unnatural flow
13- - Generic problem descriptions without specific error messages
14- - Content that seems templated or boilerplate
17+ - Use of em-dashes (—) in casual contexts
18+ - Excessive use of emoji, especially in technical discussions
19+ - Perfect grammar and punctuation in informal settings
20+ - Constructions like "it's not X - it's Y" or "X isn't just Y - it's Z"
21+ - Overly formal paragraph responses to casual questions
22+ - Enthusiastic but content-free responses ("That's incredible!", "Amazing!")
23+ - "Snappy" quips that sound clever but add little substance
24+ - Generic excitement without specific technical engagement
25+ - Perfectly structured responses that lack natural conversational flow
26+ - Responses that sound like they're trying too hard to be engaging
1527
1628 Provide your analysis in the specified JSON format.
1729 - role : user
0 commit comments