Designing an experiment for comparing user interfaces for legal formalization
summary
A prototype normative AI tool annotates legal text segments and maps them to formal logical relations; the claim is that four measurable qualities — correctness, transparency, comprehensibility, multiple interpretation support — determine usability by lawyers. The proposed experiment uses manual legal expert evaluation on coherent regulatory excerpts, asks raters to use even-numbered Likert scales, measures inter-rater agreement, and compares tools discovered via a systematic Google Scholar query; the next step is running the user study and producing a benchmark set.
AI-written summary · about
chapters (9)
- 0:00Legal UI experiment design
- 1:04Research context and motivation
- 3:31Evaluating legal formalization methods
- 8:18Simplicity versus complexity in legal UIs
- 11:56Human evaluation experiment design
- 15:52Future work and conclusion
- 16:12Q&A on participant selection
- 17:15Evaluation pitfalls and feedback
- 18:18Evaluation standards insights
related talks
25mArtificial intelligence & design: imagining the future of UX
27mMisplAIced: Good UI Needs Good UX
20mBridging the gap between AI and UI
11mEvaluating the Usability of Privacy Choice Mechanisms
01h 16mPL and HCI: Better Together
33mUX Research and Design: Crucial to development
23mUser Experience Now Includes Machine Experience
49mEmploying UX Research Methods To Create User-Informed Text