There are two "they"s involved in the conversation but generally when using someone else's stuff the more apt question is "why should I have an expectation I can use their stuff however I want without limit".

For the "they" of your provider, who is held accountable for allowing abusive traffic, the goal is to provide you outbound connectivity but to do that they also need to ensure they don't get de-peered or their network ranges blocked for hosting abusive traffic. Even for things which don't transit a 4th party there is negative incentive to let your customers abuse each other just because the addresses are reachable. This almost always results in automated systems with limited incentive for good uses of port scanning to be allowed.

For the "they" of the end system is (most likely) they didn't make the entire system available to you, just some select services for use in a certain way (e.g. loading their website). Doing that does not provide them an obligation to continuously allow all traffic received at the address to be processed and it's very likely they'll just block you entirely as another layer of defense.