Excerpt from Jesse van Rootselaar's ChatGPT interaction:

In a hallway, indoors, or a crowded classroom, the 870 [Remington shotgun] is brutal, close quarters is its playground. Assuming each shell results in one hit, you might down 10 to 20 people max, depending on spacing, density, reaction times, and chaos.

Given that these platforms are always actively reviewing and modelling your inputs, and relative metadata, its reasonable to expect that this could have been treated with some appropriate proactive measures from OpenAI. What we see instead is "consider changing something once enough people die" ; the same kind of approach we see with most safety related things.