3 ms·
There is no way they will be able to make an AI at this point that will A) Be accurate B) Work across multiple contexts C) Run efficiently on billions of mes
by bearjaws 4y ago
There is no way they will be able to make an AI at this point that will
A) Be accurate
B) Work across multiple contexts
C) Run efficiently on billions of messages
This will just result in many false positives, and unnecessary eavesdropping on employees personal conversations.
Once its revealed an organization is using this, people will quickly move all conversations to another platform, even if policy forbids that. Resulting in an even greater security risk potentially.
And as per usual, if Microsoft gets someone fired (e.g. comes in looking for money laundering, finds out the staff member is making fun of their boss), there will be no repercussions.
- zmmmmm 4y agoaccuracy isn't a strict requirement though if you accidentally fire 10% of good people you still have 90% of them left, and if that lets you fire 80% of the staff that are committing thought-crime it's probably a win.
- _carbyau_ 4y agoEven if it works out as you say, the chilling effect on who's left will be very real.
- bombcar 4y agoThat already exists in general - this is why CEOs get surrounded by yesmen. We already filter out people who aren’t smart enough to keep their mouth shut when necessary.
- zmgsabst 4y agoThat behavior has negative impacts on corporations though, from Amazon to Disney. Scaling the problem is likely to increase the damage.