Google’s Artificial Intelligence Is Becoming ‘Human-Like’ — and That Might Be a Bad Thing

The aggression, they determined, was the result of higher levels of complexity in the AI agents themselves