This is quite concerning, from the lawsuit itself [1]:
The blacklisting process was accomplished first internally at Instagram/Facebook by automated classifiers or filters, which were then submitted to a shared industry database of “hashes,” or unique digital fingerprints.
This database was and is intended to flag and remove content produced by terrorists and related “Dangerous Individuals and Organizations” to curtail the spread of terrorism and violent extremism online.
Where can one learn more about this database? Who decides what goes into this database? Is there a governance process? How about incorrectly identified items?
We are headed back into the times of "guilty, until proven innocent" versus how it should be, "innocent until proven guilty". If the scenario was the following: letting 10 criminals go free, or having 1 innocent person imprisoned, I would always choose the latter.
[1] https://digitalcommons.law.scu.edu/cgi/viewcontent.cgi?artic...
The blacklisting process was accomplished first internally at Instagram/Facebook by automated classifiers or filters, which were then submitted to a shared industry database of “hashes,” or unique digital fingerprints.
This database was and is intended to flag and remove content produced by terrorists and related “Dangerous Individuals and Organizations” to curtail the spread of terrorism and violent extremism online.
Where can one learn more about this database? Who decides what goes into this database? Is there a governance process? How about incorrectly identified items?
We are headed back into the times of "guilty, until proven innocent" versus how it should be, "innocent until proven guilty". If the scenario was the following: letting 10 criminals go free, or having 1 innocent person imprisoned, I would always choose the latter.
[1] https://digitalcommons.law.scu.edu/cgi/viewcontent.cgi?artic...