Any system that grows too fast has these kinds of problems. When it's a small intimate circle where everyone knows everyone, reputation alone can keep people in check. Once it's larger you need to invent rules and bureaucracies and structures and you will have loopholes that bad actors can more easily exploit, hiding in the crowd, than in the small version. It's the same with the Internet or computing. Security was much less of a topic when it was mostly honest academic nerds using the Internet, and the protocol designs often didn't even assume adversarial participants. Science also still runs on this assumed honesty system that worked well when it was small.

Yes, as long as (for example) you have a huge number of people desperate to claim one or more of a very small number of slots for papers in a few prestige journals then inevitably there must be either a cattle-call selection process, or blatant preference for existing insiders, or both. But things like the fancy journals' reluctance to acknowledge or care about errors in the papers they do publish really aren't at all a necessary consequence of a high submission ratio or of a rapid growth in submissions. In fact it's not even clear that they are a consequence of those pressures at all, let alone that they do anything to mitigate them. They're just things which serve the interests of insiders while in fact making the consequences of the Darwinian pressures worse.