how can we effectively regulate AI to maximize benefits and minimize risks? what should companies and gov do?