I don't know, maybe it is still possible to compete. As AIs get better at validating facts, maybe we can undo some of the poison. Maybe. I'm not entirely sure how, I think it will be different for every field. Software is the easiest. Doesn't really matter how much misinformation is out there, programs can be written and tested. Law can probably be verified against official law books. Health would be nasty though. Hard to confirm stuff without human test subjects.
That only works if the facts are written down somewhere where machines can find it. A human must taste the banana bread and leave a review before an AI can recommend it.