Should users be able to tweak AI output?
Design it like this
Make every AI result editable or adjustable; even small controls boost adoption after errors.
What the research says
People abandon an algorithm after seeing it make a mistake, even when it beats humans. But letting them adjust its output, even slightly, made them far more willing to keep using it.
“We show that people are especially averse to algorithmic forecasters after seeing them perform, even when they see them outperform a human forecaster.”
Dietvorst et al., 2015
Sources
- Algorithm aversion: People erroneously avoid algorithms after seeing them err2015 · Journal of Experimental Psychology: General · Berkeley J. Dietvorst, Joseph P. Simmons, Cade Massey
“We show that people are especially averse to algorithmic forecasters after seeing them perform, even when they see them outperform a human forecaster.”
Abstract - Overcoming Algorithm Aversion: People Will Use Imperfect Algorithms If They Can (Even Slightly) Modify Them2018 · Management Science · Berkeley J. Dietvorst, Joseph P. Simmons, Cade Massey