It seems like fraud would be really hard to prove for a policy like this.

My very first thought is - why wouldn't the policy holder make some anon accounts and "bully" themselves, lose their "job" (twitch or youtube broadcaster for example) and claim the money.