3 ms·
OP > Algorithm aversion is a product of diverse mechanisms, including ... (5) asymmetrical forgiveness, or a larger negative reaction to algorithmic error than
by gurjeet 2y ago
OP > Algorithm aversion is a product of diverse mechanisms, including ... (5) asymmetrical forgiveness, or a larger negative reaction to algorithmic error than to human error.
Related:
The legal rule that computers are presumed to be operating correctly
https://news.ycombinator.com/item?id=40052611 https://news.ycombinator.com/item?id=40052611
> In England and Wales, courts consider computers, as a matter of law, to have been working correctly unless there is evidence to the contrary. Therefore, evidence produced by computers is treated as reliable unless other evidence suggests otherwise.
- scotty79 2y agoYour belief that this rule is wrong is an example of algorithm aversion. You feel like computer systems should be judged harshly despite mistaking far more rarely than other things assessed in courts like police witness accounts or possibly even DNA evidence.
- smogcutter 2y agohttps://www.bbc.com/news/business-56718036.amp https://www.bbc.com/news/business-56718036.amp
- scotty79 2y agoOf course there are mistakes. But still fewer than with other sources of evidence. DNA evidence is also trusted by default and yet: https://www.science.org/content/article/forensics-gone-wrong-when-dna-snares-innocent https://www.science.org/content/article/forensics-gone-wrong... https://www.nbcnews.com/news/us-news/investigation-finds-colorado-dna-analyst-intentionally-manipulated-dat-rcna142541 https://www.nbcnews.com/news/us-news/investigation-finds-col...
- smogcutter 2y agoYou’re missing the point. Of course there are (enormous, life-ruining) mistakes, but that’s not it. With software, generally the only people with the means to demonstrate the software is flawed are the people in control of the software and associated data.