Preliminary experiment with 42 users found both conversational and Q&A XAI assistance let participants outperform the model on error identification, but showed no difference between the two assistance types and only modest engagement.
Why Should I Trust You?
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Can Conversational XAI Improve User Performance? An Experimental Study
Preliminary experiment with 42 users found both conversational and Q&A XAI assistance let participants outperform the model on error identification, but showed no difference between the two assistance types and only modest engagement.