Table 7 Accuracy of the detection tools (binary approach)
Tool | 01-Hum | 02-MT | 03-AI | 04-AI | 05-ManEd | 06-Para | Total | Accuracy | Rank |
|---|---|---|---|---|---|---|---|---|---|
Check For AI | 9 | 0 | 9 | 8 | 4 | 2 | 32 | 59% | 6 |
Compilatio | 8 | 9 | 8 | 8 | 5 | 2 | 40 | 74% | 2 |
Content at Scale | 9 | 9 | 0 | 0 | 0 | 0 | 18 | 33% | 14 |
Crossplag | 9 | 6 | 9 | 7 | 4 | 2 | 37 | 69% | 4 |
DetectGPT | 9 | 5 | 2 | 8 | 0 | 1 | 25 | 46% | 11 |
Go Winston | 7 | 7 | 9 | 8 | 4 | 1 | 36 | 67% | 5 |
GPT Zero | 6 | 3 | 7 | 7 | 3 | 3 | 29 | 54% | 8 |
GPT-2 Output Detector Demo | 9 | 7 | 9 | 8 | 5 | 1 | 39 | 72% | 3 |
OpenAI Text Classifier | 9 | 8 | 2 | 7 | 2 | 1 | 29 | 54% | 8 |
PlagiarismCheck | 7 | 5 | 3 | 3 | 1 | 2 | 21 | 39% | 13 |
Turnitin | 9 | 9 | 8 | 9 | 4 | 2 | 41 | 76% | 1 |
Writeful GPT Detector | 9 | 7 | 2 | 3 | 2 | 0 | 23 | 43% | 12 |
Writer | 9 | 7 | 4 | 4 | 2 | 1 | 27 | 50% | 10 |
Zero GPT | 9 | 5 | 7 | 8 | 2 | 1 | 32 | 59% | 6 |
Average | 94% | 69% | 63% | 70% | 30% | 15% |