I work in a software house where everyone uses AI. Some of them can't even write a single line of code, let alone analyze it. I was shocked when I saw their work is CTRL-A, CTRL-V into Claude and CTRL-V into the IDE without a single neuron being activated. They even ask it to summarize and generate a response for 3 lines of text in a Slack message! (Partially because they don't know what they're doing and partially because they're too lazy to think).
Well, everyone talks enthusiastically about AI, some have unrealistic expectations (thinking that it's actually intelligent, when it is not) but what bothers me is that they're indeed faster than me so sometimes I think "why am I even resisting?". Well the answer is that I love to keep my brain active and having the control of what I'm doing. Does anyone else feel kinda similar? Am I in the wrong?
P.S. Also I just want to point out that I've seen with my own eyes the deterioration of cerebral functions in people who heavily rely on AI. I'm not talking about just "forgetting how to code" but I see them losing space awareness (invading personal space, sitting like a liquid on the chair), self awareness (loudly burping, hoarding half-drank bottles of water on the desk), focus and they're easily irritable. It's multiple people behaving like that and they weren't like this before. AI is a drug.
Tell them you'll need to rip-and-replace all of the network hardware with newer gear to be AI-compatible. Then give them a realistic estimate on how much that would cost to purchase, and downtime for implementation and testing.
Then ask them what the (specific) target goal is for this project, so that you can give them a proper cost/benefit analysis, and define a progress report (in order for a goal to be achieved, it must be measurable in some way). And if they just respond with something vague like "10% increase in productivity", ask them what they think an increase in productivity looks like for network operations and how it would be measurable.
Every time they waffle on specifics, ask them again to define what the target goal is and how AI should be applied to it. If they handwave their answer, don't push them in direct conversation, but write an after-meeting email with a summary and explain that you still need a defined goal. Copy other senior employees if you think you can justify it (e.g. CFO because purchasing, CISO because network security, etc).
Avoid just saying no, as that will likely cause problems for you. After all, you are a paid employee, it is your job to help senior leadership achieve its goals. However, you are an employee, not senior leadership, so it is not your job to set targets. Your message should be that you can provide ideas if you know what the goal is.
And that goal should be SMART:
This is the way. I asked how we were tracking our AI cost benefit in a town hall and the group leadership noted that because we were delivering AI use cases on a "DIY" basis that it was on the use case owner to define the metrics. They also noted that it was typically being forgotten about on each recorded AI use case that's been floated.
If there's no specific goal then hey, no problem, pick some part of your infrastructure that you want to work on for real, mark the project as speculative/exploratory because that's all it can be with no target metrics, then spend the budget on your actual technical debt problems and call it "modernizing to support AI implementation".
Your boss will be happy with it because he can tell his boss that you're working on implementing AI (see here's the project description and the report), and you can spend time and money on solving some actual problem. Claim any benefits that come from solving the actual problem as the result of the project.