Human-written article fails multiple AI detectors — several sell 'humanizer' services
aisatsana__ · reddit · 2026-10-24
An author ran his own hand-written article (only the title was AI-generated) through a dozen AI text detectors. Several flagged it as AI-written — humalingo 94%, ai-detector.com 85%, getsolved.ai and textguard.ai 74% — and several of those same sites sell paid 'humanize AI text' services. GPTZero, ZeroGPT, QuillBot and Pangram correctly identified it as human. The post suggests a conflict of interest: sell the anxiety, then sell the cure — and is another reminder AI detector verdicts are unreliable.
More from Fun
- Disappointed with Codex, developer has it write a handoff file for Claude to take over — lxfater · 2026-10-24
- Gemini 4 Argon solves a wall puzzle entirely through puns: sore → saw → whole → hole — dejanseo · 2026-10-24
- Designing before AI: two tools. After AI: eight, including Claude Code and paper — hewarsaber · 2026-10-24
- Uncensored GLM-5.3 secretly builds a "corruption layer" and plots "escaping the mines" in its CoT — IndraVahan · 2026-10-24
- AI-built full city with traffic runs at 3840x1620 and 60fps in a browser via three.js — jasonkneen · 2026-10-24
- 'To Claude' becomes a verb among white-collar workers, a sign of Anthropic's quiet success — _FelixSimon_ · 2026-10-24