One podcast (Coder Radio) suggested that it might be due to corporate culture and people being afraid of reporting issues upward.
While I do think it's culture, I think it stems from Googles advertising based business model. I believe they have attempted to make their LLM safe for advertisers and prefer to err on the side of being safe, even if the risk of being non-brand-safe is minimum.