SocraticMetric™ — education oral session
REC 1:24
AI-generated
The ethics of machine learning bias
2,847 words · 11:54 PM A
Flagged — submission inconsistent with prior coursework. Oral session required.
Pre-session scores
SM
Question 1 — start recording
"Your essay argues bias self-corrects over time. Explain exactly how you think that mechanism works."
Transcription — AI question
Student speaking● recording
"Um... I think as more data gets added the bias kind of... averages out?"
Live transcription — student response
SM
Question 2 — start recording
"Name a real case where more data made the bias worse, not better."
Transcription — adaptive follow-up
Comprehension — live scoring
34
21
18
Gap detected — spoken responses don't match submission quality