How Do It Know?
How Do It Know?
Pay no attention to the Algorithm behind the curtain.
Originally shared by Timothy Street
“Think of this bill as an experiment in the world of algorithmic accountability, sent out much like Captain Picard, from ‘Star Trek,’ would send out a probe to explore a wormhole . . . .”
Applied algorithms of software are the product of multiple people times company policies plus user directions. Therefore, flaws, constraints, and desires are constant sources of errors.
This is not mathematics; it is social engineering - the danger should be realized and confronted.
https://www.newyorker.com/tech/elements/new-york-citys-bold-flawed-attempt-to-make-algorithms-accountable?google_editors_picks=true
https://www.newyorker.com/tech/elements/new-york-citys-bold-flawed-attempt-to-make-algorithms-accountable?google_editors_picks=true
Pay no attention to the Algorithm behind the curtain.
Originally shared by Timothy Street
“Think of this bill as an experiment in the world of algorithmic accountability, sent out much like Captain Picard, from ‘Star Trek,’ would send out a probe to explore a wormhole . . . .”
Applied algorithms of software are the product of multiple people times company policies plus user directions. Therefore, flaws, constraints, and desires are constant sources of errors.
This is not mathematics; it is social engineering - the danger should be realized and confronted.
https://www.newyorker.com/tech/elements/new-york-citys-bold-flawed-attempt-to-make-algorithms-accountable?google_editors_picks=true
https://www.newyorker.com/tech/elements/new-york-citys-bold-flawed-attempt-to-make-algorithms-accountable?google_editors_picks=true
I have lost heart that we will find the courage and the brains to confront technology. ; )
ReplyDeleteI find myself wondering if all the people who used to make these decisions could thoroughly and convincingly explain their reasoning. Algorithms are just steps to process data and reach an answer, after all, and existed in our heads long before they existed in computers.
ReplyDeleteMahlen Morris They might not be able to, or might not want to, in the cases where public servants weren't acting in the public interest. But, we, both individually, and as a society, have pretty deep experience in the misbehaviors of our fellow humans, who are most often motivated by societal pressure or self interest, or both. The Algorithm tends to be both more impartial, and more inscrutable in its errors of judgement. It's that last part that will require some thought in how to manage.
ReplyDeleteDrew McCarthy, I agree that any system of deciding, be it a human being or a piece of software, should be continually measured in a statistically meaningful way. Both code and humans have overt and hidden biases. But I suspect that asking systems to explain "why" will mainly lead people to argue over the reasons, when it should be the goals that should be the focus. Because reasons for decisions lead to arguments, whereas goals lead to solutions (or at least improvements). E.g., nytimes.com - Even Imperfect Algorithms Can Improve the Criminal Justice System
ReplyDeleteMahlen Morris Simply put, I believe you have greater faith in Algorithms than I do. I think that Algorithms generally do achieve their "goals". But the edge cases and unintended consequences can be epically bad, and tend to be unmoderated (or unmodifiable) as a result of proprietary restrictions.
ReplyDeleteNot everyone has agreed upon goals . . . . I can picture a scenario of proponents for national socialism rolling in their graves, bewailing that "if only we had had that technology" - Cassandra
ReplyDelete