How we might protect ourselves from malicious AI
by from MIT Technology Review on (#4FE77)
New research could make deep-learning models much harder to manipulate in harmful ways.New research could make deep-learning models much harder to manipulate in harmful ways.