I am interested in the potential for AI to make my work less tedious, so I regularly consume information about AI in industry. One thing I keep encountering is business leaders claiming that AI is going to make employees more efficient. In its current state, I don’t really see how.
I am probably one of the heavier AI users among the managers at my company, and I still struggle to integrate it into my day-to-day work. I use ChatGPT Plus at home, and at work we have a sandboxed AI based on an older version of GPT. ChatGPT is more usable at home because it is more capable than the LLM at work.
Even with those tools, I find myself guiding the AI every couple of minutes. This is annoying. It is like having an employee who can barely function without me hand-holding them through everything.
I’ve caught myself spending time trying to get AI to build a prompt that I can reuse to organize specific reports. I go back and forth with it trying to get the prompt right, only to realize I could have organized the report manually in less time.
That’s the part of the AI productivity discussion I think we’re missing.
For me, AI at work is still mostly a chat interface that I use as a better Google or to summarize text. It is good at both of those things, but most of my job is not searching for information or summarizing text.
There is a big difference between making a task more efficient and making someone’s work more efficient.
I think AI starts to really change the way most of us work when it becomes integrated into our actual processes. Not another chat window. Not a simple LLM bolted onto software we already use. It needs to understand enough of the process, context, and desired outcome that using it doesn’t become another task I have to manage.
Until then, there is a strange productivity trap with AI. You can spend a surprising amount of time trying to make it save you time while you babysit it