Anthropic says Smart Reports shouldn’t be used to evaluate employees, then lets management attach names to their Claude sessions. Apparently workplace surveillance becomes ethical if you put a disclaimer on it. The real innovation here is outsourcing micromanagement to the machine.
AI can make decisions and micromanage. What’s stopping the higher ups from replacing lower management now?
Kissassery, mostly. Why have one AI inflate your ego when you can have twelve department heads and one AI stroke your ego?
As I understand it this exact application of AI is currently prohibited by the EU.
They have to keep underlings around or they’re next on the chopping block
Your work email, slack, google drive, or any other account given to you by your job should not be expected to be private.
There’s a small legal carve out for health data through your employer health insurance and one or two others probably, but its well within your company’s rights to see the communications that are done through company resources.
Micromanagement to a whole new level.
Is it really any different than companies having full access to your email, teams chats, and other correspondence tools provided by the company?
The only thing protecting us before was “who has the time to analyze all the data we gather about you?”.
Without analysis how could they tell that you consistency take an extra 1.45 minutes to clock back in from lunch when you eat Taco Bell. Schedule a colon cancer screening, while also increasing your health insurance premium along with your estimated new risk level due to your sub optimal life choices? This is all needed to insure you last meat bags with jobs perform to shareholder expectations.
Well guess what LLMs are really good at?
Pretty sure you can‘t do that in the EU. Your boss is not even allowed to give AI details about you or consult it about you. The thought that Anthropic wants Claude to become your supervisor is deranged and dystopian.
I am so done with the software industry
There’s a pile of products that already monitor LLM use, prompts, data sent, etc. There’s a whole branch of cyber called AISPM tnat deals with it.
AI manager is just a manager.
