I'm not hurt by words like that. I was thinking in general terms of moderation, slurs is one, but simple automated spam would be another thing, I imagine it isn't protected against. I just thought it was funny to see my initial concern verified so quickly.
I read the description, and thought, this sounds like a bad idea. Where is the content moderation?
I scrolled down, and the first post I see is the n-word
The food example, is the exact reason for large farming subsidies in the European Union. These were implemented as a founding initiative, due to the experience of food shortages during the second world war.
A great number of things could be considered critical. Due to the nature of when access could be cut off, the main thing countries likely worry about being able to access, is things humans need to stay alive, and things needed to wage war.
They did not call phishing, but their point still stands. A phishing email is malicious, and if you see this kind of prompt injection as malicious, then I don't think it's a stretch to call software that engages in malicious prompt injectic malware