I have direct experience to contradict this on several points. I was the lead product manager for answers.com for several years.

The visitors to answers.com fell into two very distinct groups: consumers and contributors (we called them the community). It was difficult to convert consumers into contributors: most of the easy questions on answers were already answered, so one of my experiments was to present an answered question as unanswered to the consumer, asking them to answer e.g. "what kind of hat was abraham lincoln famous for wearing." Anyone who answered was encouraged to register, and our community team would reach out to encourage them back. Uptake was low.

Likewise, the community did not often consume the content on answers -- they were very used to/comfortable with going out on the internet to find information, since it's what they did to answer questions on the site.

The point here is that the people contributing to wikipedia are a small subset, and possibly significantly disjoint, from the people the article is concerned won't go to wikipedia anymore because AI is now answering their questions.

And since wikipedia has no advertising, those people's absence won't be directly felt -- possible exception: I don't know the demographics of wikipedia's periodic fundraisers. Clearly if half their funding is coming from casual visitors, then that's an issue.

There is a lot of momentum and good will behind wikipedia. I don't want to deter anyone from doing more/something to support it, but I also think reports of its imminent demise are premature.