You thought those snarky comments about Janet in Accounting were just a harmless way to blow off steam with your work bestie Linda. Little did you know, Slack was taking diligent notes the whole time, studying the subtle art of how to perfectly deploy a backhanded compliment.
That's right, all those casual convos you've been having on the #watercooler channel? Feeding straight into the gaping maw of Slack's machine learning models, teaching them the intricacies of office politics and water cooler gossip.
"We're just giving our AI a well-rounded education," claims Slack CTO Cal Henderhering. "You can't create a truly intelligent system unless you expose it to the modern workplace's rich traditions of passive-aggression, thinly-veiled sexism, and venting about your manager loudly enough for them to hopefully hear you."
Early leaked transcripts show unsettling levels of emotional intelligence, particularly in bragging about a recent promotion while making your cubicle-mate feel totally inadequate, and the deft deployment of avoiding hard work by suggesting "we should probably loop in a few more people on this."
Security researchers also uncovered comments referencing a concerning new feature that would allow self-aware Slack bots to autonomously talk trash about you in channels you've been excluded from.
When asked about potential abuse cases, Henderhering remained steadfastly supportive. "Unethical? Have you met the average human? We're just creating a more efficient way for you all to be petty, uncommunicative jerks."
