There are many, many problems with the almost cult-like atmosphere of reviewers in good journals but pay-to-play isn't one of them.
If we're going to beat that kind of misinformation, the real stuff needs to be open and available so that translators can help people understand it. There were people in academia doing this on social media during the pandemic, but we need many, many more of them. There are a lot of people outside of academia who have enough of a background that they could help - if only they had access.
Edit to add: Also, I've heard from some folks in academia that they've been bitten by open access journals that started out well intentioned and turned into pay to play with bad reputations. And that in some fields, it is hard to track which ones are good and which ones are going bad.
Expecting it to be well explained to the general public is like asking all patches posted to LKML to include a lay-language explanation of exactly what the patch changes, without using specialized CS terms.
Ok, but I take some issue with this.
Most scientific papers are not intended at the general public, but at experts working in the specific area that the paper covers. Also, often these papers are not even understandable by a phd outside the area, let alone a lay person.
Which journals and authors are reputable are known to experts in the area. "Pay to play" papers published in obscure journals will not get much attention or be taken seriously.