I have a lot of issues with Google as a company, but I do not envy their position when it comes to the web search. It is a cursed problem no matter how you look at it.
When you search for "vacuum cleaner", you either want online stores or you may want a local store. In either case it's a business query rather than looking for reviews or specs.
What's really polluting search, including Google, is the on-demand content generation based on your query. It's a sea of flotsam.
Look at any two sided market: Amazon/Etsy's third party sellers also want to be at the top, and there are bad actors trying to scam the intermediary and the customer. The same low friction onboarding that allows the company to succeed as an intermediary is also facilitating the fraud. Merchants can be fraudsters, advertisers can be fraudsters, users can be fraudsters producing fake clicks. There's economic incentives everywhere. See the relatively recent news of spotify getting generated songs, which are then played by fake listeners.
The companies are entering this situations with their eyes wide open. It's a necessary problem when you want to be an intermediary at a large enough scale. And if you don't have the scale, the per-interaction costs are really high, and your company gets beaten by someone with an easier onboarding funnel.
But the funding model is also extremely blessed: you have a stream of people telling you every page load what they're interested in at that exact moment, and it's not hard to match them up with advertisers for those keywords.
And arguably the SEO parasitism problem is worsened by their monopoly -- to the extent that optimizing a site for one engine deoptimizes it on on others, the stronger the lead engine #1 has the more incentive there is to game it.
How many millions do you need to spend to decrease your own spending on AI for every query?