Addressing the post, a focus on AI risk feels like something worth experimenting with.
My lame model suggests that the main downside is that it risks the brand. If so, experimenting with AI risk in the CFAR context seems like a potentially high value avenue of exploration, and brand damage can be mitigated.