Hmm, you could work on an open-source project for detecting and/or mitigating corporate and government influence campaigns in the fediverse. I think, right now, the fediverse is still too small for it to make sense for those actors to deploy expensive, sophisticated attacks against it, but if the community grows that threat is going to grow with it, probably at least as long as we're alive. I assume you're talking about financial fraud; I don't know how similar detecting "fraud" accounts would be but I imagine there's some overlap.

I think it would be fun to work on a bot / influencer detector. And if it works, it would hopefully make the fediverse more resilient to coordinated influence campaigns designed to destroy or capture it. Which would help to ruin the corporate media hellscape that has been wreaking havoc on the world ever since the inception of feed manipulation.