Maybe if their rate of infection were significantly lower than the local population, you could argue that we would be decreasing the percentage of the infected

Maybe if their rate of infection were significantly lower than the local population, you could argue that we would be decreasing the percentage of the infected population which doesn't feel particularly compelling to me... but we'd really need to know how much each group contributes to the spread. My expectations is that illegal immigrants make larger contributions than average given their density in living arrangements, lack of access to health care, inability to take sick days, etc. I think we are only in the business of arguing the counterfactual of what percentage of them get covid from interactions with law enforcement vs otherwise.