Thread by @Smerity: "Today's meta-Twitter summary for machine learning: None of us have any consensus on what we're doing when it comes to responsible disclosure […]"

TECH PLAUSIBLE DENIABILITY Facebook and Twitter "algorithms" are MACHINE LEARNING (ML) models. But they use PEOPLE to train their models. Like a kid that "learns", these are trained. If/when Congress holds hearing on tech bias you will hear a lot about how these companies "do not target conservatives". What they aren't going to tell you is they ARE, but it's the ML models doing their thing, making decisions ON THEIR OWN based on what they've been taught.