That assumes the chatbot creators don't introduce explicit bias of their own. What if they start doing that? For example a search for nutritional information about some foods might be influenced by fast-food companies buying "ads" to insert their own messaging.