I see, then ignoring the reactions of an organism, how can you tell whether it is aware or not?
This is the domain of the Turing test.
The test strategy envisaged by Alan Turing is not really good enough now, there are already machines that can almost pass as sentient in blinded conversations. Here is Eliza, a virtual therapist you can chat to if you have problems :
http://nlp-addiction.com/eliza/Testing for awareness needs to become far better in the future because we will need to know if the machines we build based on neural architecture models really do develop inner sentience or just give the outward appearance of being conscious. The testing for consciousness is likely to be based on a combination of machine learning and imaging technologies. We are already started down this road, programming computers to learn recognise tell tale patterns in activity in brain imaging scans that correspond to various arousal states - vegetative, coma, asleep, awake, just waking up, just going to sleep and so forth. By developing and adapting such technologies we hope to be able to develop a reliable test for genuine inner experience in a machine.