As tech moguls and politicians weigh in on AI safety concerns, the question remains: can superintelligence truly be kept 'under human control'? Microsoft's new code of conduct aims to address these concerns, but critics argue it may be too little, too late.