Listicles are hacky, I know. But in any case, here are ten AI safety projects I'm pretty into.
> AI tools for fact-checking
Nathan Young has an MVP extension for this.
> 3. Tracking sketchy AI agent behaviour “in the wild”
I have an MVP of this here https://docs.google.com/document/d/18e07i1-L6K4S12eLn-AwJVdAHoqBVQ4adq_n_vWGlWM/edit?tab=t.0#heading=h.67vn92etlqtm / https://github.com/NunoSempere/ai-osint-mvp
tl;dr: The individual twitter accounts I was tracking were impish but ultimately harmless.
> AI economic impacts tracker
Dan Carey might be looking for some funding to do a macroai group, but not sure if it's similar to what you are thinking of.
> AI tools for fact-checking
Nathan Young has an MVP extension for this.
> 3. Tracking sketchy AI agent behaviour “in the wild”
I have an MVP of this here https://docs.google.com/document/d/18e07i1-L6K4S12eLn-AwJVdAHoqBVQ4adq_n_vWGlWM/edit?tab=t.0#heading=h.67vn92etlqtm / https://github.com/NunoSempere/ai-osint-mvp
tl;dr: The individual twitter accounts I was tracking were impish but ultimately harmless.
> AI economic impacts tracker
Dan Carey might be looking for some funding to do a macroai group, but not sure if it's similar to what you are thinking of.