The obvious on is AI engineer, dataset creator or manager. Regardless of how advanced the technology gets, it will always require humans to oversee how it gets implemented in order to solve human problems. Some of these require relatively in-depth knowledge of the things that AI will recreate such as art or copywriting. So fields like visual arts or illustration will still be viable career paths but anyone going into there must be 100% aware of working with instead of competing with AI.
Medicine is also a field that will always require human doctors. AI might be able to diagnose and treat but it's hugely unlikely that a machine will be allowed a medical licence. AI is already huge in medical fields such as assisting surgeons or dentists in planning complex cases in a fraction of the time, and thus increasing their revenue. The humans who earn the right to charge for using medical AI are those who will cash out.
International relations are also relatively safe from AI. Specific areas like interpretation or translation may be entirely replaced but you can't outsource diplomacy between real countries to a machine. Similarly, all services designed to support those industries would be relatively safe. People still need food, transport, hospital care, entertainment, etc.
The biggest thing to bear in mind is children need to grow up embracing AI rather than rejecting it. If you get them started early with the basics (learning prompts, LLM, generative AI) then it will be much easier for them to find a niche later one. Do not encourage a blanket rejection of AI which some people are doing right now. Yes there are potential downsides but none of those issues are fully factually confirmed or set in stone right now. Whether the anti-AI people like it or not, the technology is coming and no amount of moaning on social media will change it.
Boycotting AI right now is basically the most suicidal act someone could do and will condemn themselves and their children to poverty and misery in the next few decades.