How to Detect AI Cheating in a Paralegal Case Study Interview
Paralegal case studies typically present a hypothetical filing deadline conflict or a document-review scenario. AI cheating shows up as a candidate reciting an AI-drafted legal-process answer that's fluent but generic, unable to adapt to the specific jurisdiction or court rule you name. Neuroxa's AI Meeting Proctor tracks gaze and window activity through the live call.
| Threat Model | Observable Tell | Confidence |
|---|---|---|
| LLM window generating a legal-process answer in real time | Fixed-point gaze before a fluent, generically "correct" answer; can't adapt when you specify a jurisdiction | High |
| Memorized generic answer regardless of your specific fact pattern | Answer ignores case-specific facts you gave (deadline, court, document type) | Medium-High |
| Screen-share/second-monitor reveals a legal-research tool or AI assistant | Tab/window-focus change logged during a "thinking" pause | High |
| Second person feeding answers via chat | Audio-lip sync delta exceeds natural conversational range | Medium |
Interviewer script: "Here's the scenario: a motion is due in [specific court] in 48 hours and the attorney is unreachable — walk me through your next three actions." Change one detail mid-answer (e.g., swap the court or add a conflict) and see if reasoning updates live.
Evidence to capture:
- Full call recording with gaze overlay
- Tab/window-focus change log
- Response specificity against the fact pattern you provided
- Audio-lip sync analysis
- Follow-up adaptability when a detail is changed
Neuroxa product: AI Meeting Proctor — live gaze and screen-share monitoring for scenario-based legal case study interviews.
FAQs
Is it fair to expect jurisdiction-specific knowledge from every paralegal candidate? No — the point isn't testing legal-research recall, it's checking whether the reasoning process is genuinely the candidate's own versus a recited AI script.
What if the candidate references a real case they worked on? That's a strong signal of authenticity — cross-check for internal consistency with their resume, which AI-generated hypotheticals typically lack.
Should we disallow legal-research tools during the case study? Yes for the live reasoning portion — legitimate research tool use is more appropriate for take-home assignments with disclosed AI-use policies.
How do we handle candidates who freeze without their notes? Some hesitation is normal; the flag is fluent recitation combined with inability to adapt, not hesitation itself.
Related: Paralegal Zoom Panel Interview · Paralegal Phone Screen · Paralegal Contractor/Offshore Onboarding Verification · Accountant Case Study Interview
Run genuine legal case-study interviews with Neuroxa.ai AI Meeting Proctor.