If you order through e.g. instacart somebody usually still has to make an in-person trip to the store and take things off shelves. It's the same vector, but you are paying someone else to be exposed to it.
It's not necessarily true but it is statistically true. Gig workers are more likely to be non-white and have lower educational attainment than traditional workers, and non-white gig workers are more likely to rely on gigs for primary income than white workers.
Overall you could posit that a single gig worker doing the shopping for 20 households is safer for everyone that 20 individuals doing the shopping for 20 households. The gig workers are often shopping for multiple clients at once - fewer trips, fewer people, fewer chances of infection.
The delivery person could have a higher risk due to visiting all the households, but my experience is that they are all hands off. The gig worker drops the groceries at the door and leaves.
It certainly reduces risk for the clients, but also increases risk for the gig worker. It might reduce societal risk overall, though I'd argue it's hard to say without a deeper understanding of overall behavior, especially since poor people are already at higher risk due to a number of other factors.
Concentrating harm in an already disadvantaged group in the name of the 'greater good' is ethically thorny though. It can be used to justify a lot of not-so-great things.
There are more people involved here than the gig workers and the advantaged, The grocery store employees are at a decreased risk if one person does the shopping of. The managers may have a pretty comfortable salary but the bag 'boys' or the workers stacking pallets behind the scenes could also be disadvantaged.
If a gig worker reguarly goes to multiple grocery stores (which is pretty common, moreso than for individual shoppers), there might not be much benefit for grocery store workers. The ability to transfer disease between grocery stores might increase risk in a major way. It's also not clear that interacting with one high-risk gig worker is better than interacting with twenty low-risk remote workers.
I'd be wary of coming to strong conclusions here without data; you can easily come up with plausible arguments either way.
It's not necessarily true but it is statistically true. Gig workers are more likely to be non-white and have lower educational attainment than traditional workers, and non-white gig workers are more likely to rely on gigs for primary income than white workers.