This is a valid concern and must be considered in applying and configuring the crosswind air barrier into specific environments like a restaurant. You must be referring to a Korean study that found customers who were in line with direct airflow from the air conditioner did get infected. However, just as interestingly, others outside the flow but as close to the index case, did not. So this is evidence that airflow can indeed carry the infection for several meters, while on the other hand this study demonstrates support that the crosswind air barrier has good mitigation promise in restaurants.
Why? It is unlikely that the Korean air conditioner had an effective high-MERV air filter so it recycled infected air back to the customers behaving like a regular fan. Instead of decreasing the viral load with time, it increased the load, throwing the ever-increasing viral load at the healthy customers. In a permanent hands-free installation crosswind units in a restaurant, the units can indeed coexist with the A/C and be very beneficial (see also Question Q3). Multiple crosswind devices can be ceiling-mounted where their air streams are pointed downward towards the floor hitting your customers, while A/C units should have their exit ports pointed away from your customers.