Both companies plan to invest at least $1 billion each in AI safety over five years
Automated researchers outperformed human proposals while improving 10 safety failures
