That kinda makes the "racist assumption" (no personal offense) that you could neatly divide people into different groups that would be internally homogeneous and externally distinguishable. So far any generalized attempt to do that has failed and the failure often let to catastrophic results.
It's almost physically impossible for a group (2 or more individuals) to be fully homogeneous and that impossibility increases with group size and complexity of it's members.
Assume for a second there is an algorithm (some sort of step wise tutorial on "How to ..."). Doesn't matter what this algorithm does, could be language, cuisine, how you build houses, how you express emotions, how you make friends, how you are polite, ... or whatnot. Now assume that algorithm works 99% of the time. And because that's fairly good EVERYBODY is doing that.
Now for the first time and for the first 10 people that might work perfectly, but either over time or because the group size increases you approach the law of large numbers where, on average, for in 1 out of 100 cases, it fails.
Now how would you rate that failure? Is that person for whom it fails from a different group? Did a different group just emerged from the different result? Is it mixing of groups?
Either way the membership of the group does not tell you how any individual will perform on that. Meaning in return the membership of the group is not a general description of the individual. So either the algorithm must be dumped down so that it works 100% which likely comes at the price of it's usefulness/versatility or you need to narrow down the descriptive power of group membership, so that they would only share one attribute/ability/behavior making it rather superficial. You could also decrease the complexity of the members, but that likely is received as torture.
While the more attributes you want to combine the smaller the group gets to the point where it makes more sense to think of them as individuals.
And as a consequence of that every larger group is essentially a conglomerate of subcultures and every individual is a member of a multitude of different groups, none of which individually would describe them in general while being meaningful.
So depending on the feature that you're distinguishing, the idea of "mixing" might make no sense to begin with because the group is vastly heterogeneous in the first place. So it would already be "mixed".
Which makes this narrative of "mixing" a quite insidious one, because it's not really about introducing something, but it's about dealing with something that is already there or rather the attempt to NOT deal with it, by alienating the problem and the person having that problem even or especially if that is not the person's fault but rather a symptom of failure of a popular narrative or cultural algorithm. Rather than admitting it only works 99% of the time and working on alternatives for when it doesn't, the narrative is that the exceptions to the rule just do it wrong/are different are themselves the problem.
So it's not really about "mixing of groups" it's about broadening your circle of compassion in terms of who's in your group and who's perspective also matters or whom you can ignore as "others". So you don't really need an external group for there to be conflict and war you just need to have an internal situation that is not seen as a problem by some and seen as unbearable by others.
Now getting into contact with people with different experiences and listening to their accounts and trying to understand their perspectives might help in getting around misconceptions but that doesn't mean that there aren't still conflicts within society and the wider human community. Whether there's war or peace depends on whether you can and want to solve these issues and knowledge about them is necessary but not sufficient alone.