A major transnational audit of conversational AI assistants by public broadcasters has delivered a stark verdict: widely used chat systems are producing unreliable news answers at scale, with nearly half of sampled responses containing at least one significant problem — a result that should...