As someone who works in a knowledge industry, anyone relying on AI for their workload will end up with more errors than solutions. IT requires a high degree of accuracy in the information you handle that gets you to a solution. Out of everything you can say about AI, you can’t say that it’s highly accurate.
Any time I’ve given a technical question to copilot or chat GPT, I usually get nonsense back. It will not help me solve any of the issues I need to solve as a part of my job.
I understand how the current version of “AI” works, and from that knowledge, I know that for any meaningful task I face with even a small amount of complexity, these so-called “AI” bots can’t possibly have any relevant answers. Most of the time I can’t find relevant answers on the Internet by trying. Sometimes I only get adjacent information that helps lead me to the unique solution I need to implement.
“AI” in IT support actually makes things go slower and cause more issues and frustration than actual tangible help with anything that needs to be done. You end up going down rabbit holes of misinformation, wasting hours of time trying to make an ineffective “solution” work, just because some “AI” chatbot sent you on a wild goose chase.











Machine learning for OCR makes a ton of sense. Human writing is highly dynamic, especially handwriting. It makes sense that OCR would benefit from a trained model for recognising words.