Read on Twitter
We've written several times about what we describe as Phrenology 2.0 — the attempt to rehabilitate long-discredited pseudoscientific ideas linking physiognomy to moral character — using the trappings of machine learning and artificial intelligence.
For example,, we've put together case studies on a paper about criminal detection from facial photographs... https://callingbullshit.org/case_studies/case_study_criminal_machine_learning.html
...and on another paper about detection of sexual orientation from facial structure.
(tl;dr — both are total bullshit) https://callingbullshit.org/case_studies/case_study_ml_sexual_orientation.html
(tl;dr — both are total bullshit) https://callingbullshit.org/case_studies/case_study_ml_sexual_orientation.html
A recent Harrisburg University press release, abruptly removed from the press office website, promised another entry into the phrenology 2.0 canon. https://twitter.com/CT_Bergstrom/status/1258158701052977152
Now @SpringerNature is planning to publish the Harrisburg work.
Irrespective of whether you think publishers should be responsible for not publishing harmful pseudoscience that furthers social injustice, the Coalition for Critical Technology @forcriticaltech has put together an exquisitely researched explanation of why this work is bullshit.
Their article makes our previous writings on this subject look like two guys ranting on the internet. If you care about #algorithmicinjustice, you need to read this. https://medium.com/@CoalitionForCriticalTechnology/abolish-the-techtoprisonpipeline-9b5b14366b16
Why does it matter? Would anyone ever be stupid enough to use this sort of shit in a policing context?
Sadly, we're already seeing horrific misuse of machine learning technology in policing. This story was published today. https://www.nytimes.com/2020/06/24/technology/facial-recognition-arrest.html?smid=tw-share
Sadly, we're already seeing horrific misuse of machine learning technology in policing. This story was published today. https://www.nytimes.com/2020/06/24/technology/facial-recognition-arrest.html?smid=tw-share
Even if the algorithm itself weren't racist garbage, the approach of dragnetting for suspects using facial recognition technology is guaranteed to generate false positives at a very high rate.
I explain in this thread: https://twitter.com/CT_Bergstrom/status/1275809222240792576
I explain in this thread: https://twitter.com/CT_Bergstrom/status/1275809222240792576
"No one is getting convicted based on this technology alone," one might argue.
Not yet.
But I don't give a shit. People are getting arrested. Arrest is always traumatic. And for black men, arrest is all too often fatal; when not fatal it too often leads to false convictions.
Not yet.
But I don't give a shit. People are getting arrested. Arrest is always traumatic. And for black men, arrest is all too often fatal; when not fatal it too often leads to false convictions.
People who understand the limitations of technologies need to take a stand.
We need to explain why they are injust—and bullshit as well.
We need to work to halt the use of these technologies.
We need to ensure that ML is not used in service of injustice.
We need to explain why they are injust—and bullshit as well.
We need to work to halt the use of these technologies.
We need to ensure that ML is not used in service of injustice.
