I'd like to help with AI safety, but after reading the other articles, I'm still not clear on how. How can I figure this out?

Here are some things you could try: