"If an algorithm can help send someone to jail, shouldn't we check that it actually works?"

[CAPT_Irrelevant]

The heading has been taken from this tweet.

New Scientist (newscientist)

The linked article is paywalled 4 paragraphs in, cutting off at the mention/hyperlinking of an application called “COMPAS Classification” that the article is talking about the nature of (the article cuts off half of the name, but google helped get around that).

Classification - equivant

This application advertises the following:

The COMPAS Classification software is a management support tool that can be interfaced with existing jail management systems to provide critical inmate management information to help support all aspects of a jail’s classification operations.

Features found in the COMPAS Classification software include:

Classification – Decision Tree Primary Classification and Reclassification Instruments
NIJ Mental Health Screening & Brief Jail Mental Health Screening
PREA Assessment and After-Action Plan
Inmate Housing Management
Inmate Discipline
Gang Tracking
Inmate Programs Management
Inmate Complaint/Grievance Management
Reporting

Obviously this is quite central to the American legal/prison system, but would such a thing hold any benefit to our own?

There would also be the matter of what the article insinuates, which makes me then put this application in the same category as the infamous “Turnitin” which for one of the reasons it’s controversial is due to its closed-source nature.

[tjc (Tim Challis)]

Automated judge, jury & executioner? Just what Centrelink wants for a RoboDebt upgrade, stat!

You just know funds will be available for this. Security theatre buys right-wing votes wholesale.