Microsoft Purview: Additional classifiers for Communication Compliance (preview)

Anyone using Teams is already a red flag.

Many workforces in Australia are highly unionised. Unions have been largely effective - even in recent years - at using their collective bargaining power for the good of the worker. Hell, our ruling political party is literally called the Labor party. It has strong union ties and a history of passing pro-worker legislation. Union corruption exists as it exists in all areas where power can be had, but in Australia’s case you’d truly be throwing the baby out with the bathwater by saying “unions are bad!”, even in industries like tech where we’ve never really sought collective bargaining on a large scale, the universal protections ushered in by the union movement benefit all workers here. It sounds like you’ve fallen for the same propaganda as everyone else but you particularly think that you have “smarter” reasons. You don’t.

Pretty clear one of the major things they're going for here is detecting "jobsite troublemakers", ie employees who are upset with job conditions/agitating for improvements/discussing salaries/etc, which is given specific legal protection. It is explicitly legal and protected for employees to discuss labor conditions, organizing, or salaries regardless of whether you do it "on company property" or "on company chat". Just because the company owns it doesn't mean you have no legal rights - just like a company can dismiss you for no reason but they can't dismiss you for any reason.

They are wrapping it up with "think of the children" justifications like "employees who are discussing salary might be considering leaving and they might take nefarious action if they do so" but that's the core of the situation here - these are tools to detect and fight against legally-protected activities by employees.

> Workplace collusion: The workplace collusion classifier detects messages referencing secretive actions such as concealing information or covering instances of a private conversation, interaction, or information.

> "The leavers classifier detects messages that explicitly express intent to leave the organization, which is an early signal that may put the organization at risk of malicious or inadvertent data exfiltration upon departure"

Hypothetically, do you think it would be a good idea for Microsoft to build a classifier and provide managers with a list of potential "religiously devout", eg based on correlated work/away periods, language patterns, etc? Sure, it's a legally protected classification, but there's an elevated risk of extremist activity, which surely presents a business risk, right? So why not?