That $75k number quickly balloons up to quote exorbitant amounts if health of those social workers gets put in danger due to such interactions (whether due to getting infected or due to violence inflicted upon them).
Given that we are trying to reduce infections in the overall population by any means possible, this seems like a rather cheap and non-invasive (aka no restrictions imposed on the population) way of achieving that. Sure, masks help, but if you are surrounded by tons of infected people on a daily basis in a confined environment, you have a rather notable risk (as evident by all the cases coming from doctors who got infected after working with covid patients during the first two waves).
Why should we put lives of 1-2 social workers at risk and deal with them potentially costing way over $75k, instead of spending those $75k on a solution that has none of those risks?
If the person has symptoms, they need a swab. If they don’t have symptoms and you’re suspicious of infection (eg they are a close contact), they need a swab. The robot can’t do that.
Also, does the dog cost $75k and that’s it? Or is there ongoing costs, upgrades and running repairs? It might not need sick days, weekends or vacations, but it can’t do a fraction of the work a human can.
Bringing in a semi-autonomous robot to work amongst the public probably doesn't and shouldn't happen in a financial void.