100%!
-
-
[email protected]replied to mekka okereke :verified: last edited by
@[email protected]
It should be unsurprising that an AI trained on "these are the resumes of people we've hired and these are the resumes of the people we didn't find suitable to hire" would tend to perpetuate previous hiring-biases. -
mekka okereke :verified:replied to [email protected] last edited by
'On two occasions I have been asked, – "Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?" ... I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question'
- Charles Babbage, 1864
"Pray Mrs. Wilson and Dr. Caliskan, if you put into the machine resumes and biased hiring decisions, will an unbiased hiring model come out?" No. Y'all tripping. BFFR.'
- Mrs. Kyra Wilson and Dr. Aylin Caliskan, 2024
-
Dana Friedreplied to mekka okereke :verified: last edited by
@mekkaokereke didn't Amazon try this - and immediately shut it down for the same reason - like a decade ago?
-
@tess @mekkaokereke It wasn't even 10 years ago
https://www.bbc.com/news/technology-45809919 -
Jordan Biserkovreplied to Jenniferplusplus last edited by
@jenniferplusplus @tess @mekkaokereke
Not to be pedantic, but it was exactly a decade ago. From the article:
"Reuters spoke to five members of the team who developed the machine learning tool in 2014"."By 2015, it was clear that the system was not rating candidates in a gender-neutral way because it was built on data accumulated from CVs submitted to the firm mostly from males".
-
Jenniferplusplusreplied to Jordan Biserkov last edited by
@jbiserkov @tess @mekkaokereke Sure, 10 years since they developed the resume screening AI, for whatever that means. 9 since the developers were convinced enough that it was so irredeemably sexist that they risked highlighting it to their management. 8 since Amazon stopped using it. 6 since they caught a bunch of bad press for it.
And perpetually 0 days since some VC funded startup and/or 100 billion+ dollar tech giant thought they should change nothing and try again.
-
Shannon Clarkreplied to mekka okereke :verified: last edited by
@mekkaokereke the argument from companies like Salesforce that the models they release for academic study aren’t the ones they make available commercially which they claim to have tuned/tweaked to remove biases etc should raise all kinds of flags (for academics but also for those firms commercial customers who are then relying on smoke and mirrors and vapor
