whats up soylentils. any good meta discussion of soylent we care to have?
-- other than actual doxxing or CSAM etc don't flag posts here thx --
Reply to: Re:Site still loads slowly sometimes
(Score: 0) by Anonymous Coward on Saturday October 18, @01:48PM
by Anonymous Coward
on Saturday October 18, @01:48PM (#1421157)
Sure, but there have been mistakes and I was specifically bringing up automated methods. The world is increasingly handing responsibility to AI systems and at some point something bad will happen and people will shrug and blame the computer. It will be 10x worse than corporate BS.
As for your exact point there have been users that were falsely accused and the defense was it sounded like a banned user. That is subjective content discrimination and referring to a banned user shows how easily bias can enter the equation.
My point was more about society and how allowing AI to make these subjective decisions makes a black box method of discrimination that allows malicious abuse to be blamed on software "bugs." The banned user situation here is just an example of how subjective decisions will have errors even with a human making the calls, and I wanted to tie that into a larger point about the perils of automated moderation of humans.
(Score: 0) by Anonymous Coward on Saturday October 18, @01:48PM
Sure, but there have been mistakes and I was specifically bringing up automated methods. The world is increasingly handing responsibility to AI systems and at some point something bad will happen and people will shrug and blame the computer. It will be 10x worse than corporate BS.
As for your exact point there have been users that were falsely accused and the defense was it sounded like a banned user. That is subjective content discrimination and referring to a banned user shows how easily bias can enter the equation.
My point was more about society and how allowing AI to make these subjective decisions makes a black box method of discrimination that allows malicious abuse to be blamed on software "bugs." The banned user situation here is just an example of how subjective decisions will have errors even with a human making the calls, and I wanted to tie that into a larger point about the perils of automated moderation of humans.