In designing robots with high level of autonomy for safe and ethically acceptable interaction with humans, engineers need to have a profound understanding of human behaviour control including social interaction. The present paper was inspired by Janus, a god of transition in Roman mythology whose double face could be used as a metaphor to provide insights into current difficulties and ethical concerns of designing artificial autonomous agents. A critical review of biology and the theoretical background of ethical concerns in biology are used in order to gain insights in recent developments in robotic research and ethical concerns in the context of the history of Western Science. Aristotelian organic science was contrasted with the Newtonian-Cartesian science to discuss the dilemmas and paradoxes of robot ethics arising from insufficient understanding of the autonomy of an agency. These are presented from the two perspectives Janus head is facing: looking backwards to see the dangers and benefits of past and current limitations of our knowledge in robot designs; and looking forward benefitting from or abusing of our future deeper understanding of the principles of normal functioning of autonomous agents.