Americans are skeptical of whether computer algorithms should be tasked with making decisions that have life-altering consequences for humans, according to a Pew Research Center report.
The Pew Research Center surveyed 4,594 adults from May 29 to June 11 for its poll on public perceptions of how algorithms may be used in various real-life situations. The poll was conducted as part of the American Trends Panel, which is designed to be representative of the U.S. population.
The margin of sampling error for all respondents is plus or minus 2.4 percentage points.
Although public attitudes toward algorithmic decision-making varied by context, Pew found Americans generally question whether algorithms are fair to people being evaluated.
Most respondents agreed it is "unacceptable" to use algorithms to make create personal finance scores using consumer data (68 percent), conduct automated video analysis of job interviews (67 percent), perform resume screening of job applications (57 percent) or assess criminal risk for people up for parole (56 percent).
Pew identified four key themes driving these concerns based on responses to open-ended questions included in the survey. Namely, respondents who found the above uses for algorithms unacceptable tended to cite issues related to privacy, fairness to those being evaluated, the "human element" needed in important decisions and whether automated systems are capable of "nuance" in real-life scenarios.
To read Pew Research Center's report, click here.