People increasingly write “AI” on their CVs. I have started wondering what they mean by it.
If you have chatted with ChatGPT, then technically you have used AI. If it has become your helpful new friend, fine. But that does not, by itself, mean you know how to use AI for work. The chat interface is the front door. It is not the whole house.
You begin acquiring practical experience when you use AI to do something productive for which you remain responsible:
- as a research assistant for your PhD;
- as a development partner while writing software;
- as an analysis system for production data or weaknesses in your operations;
- as an artistic partner for creating visuals or motion pictures;
- as a way of managing work that you would otherwise fail to manage; or
- as part of your marketing analysis and the delivery of actual marketing projects.
That is when the friendly conversation becomes a working relationship—and the awkward questions begin. Which model actually suits this job? How much context does it need? How many tokens will the project consume, and what will that cost? Should you use OpenAI or xAI, one of the other Western providers, or perhaps a Chinese model that fits the task better? Can the tool plan the work, preserve that plan, and show you when it has wandered away from it?
If you have never used the planning functions in your chosen tool, compared models, watched a token budget, or abandoned one model because another fitted the job better, then you have barely started.
The most important lesson is that an AI answer is not the final answer. AI gives you options, drafts, analyses, and possible routes. It can also confidently invent things, miss the point, and sprint down a rabbit hole unless you control it. You have to set the scope, demand evidence, establish checkpoints, build guardrails, and know when to stop it.
So please do not put “AI” on your CV simply because ChatGPT has become your new helpful friend. Put it there when you can explain what you used AI to accomplish, which tools and models you chose, how you verified the work, where the system failed, and how you controlled its cost, scope, and risk.
That is experience. Everything before it is acquaintance. Acquaintance can be useful, but it is not the same thing.