Show of Hands

If an artificial intelligence is developed that shows some inconclusive signs of self awareness, should we try to develop it further or stop since it may be dangerous to go on?