Toward responsible artificial intelligence in medicine: Reflections from the Australian epilepsy project