A useful test to see the value of definitions, is to check if there are simple programs which become conscious according to the definition. There are very simple programs (<1000 lines) which do cost optimization, interact with the os and other programs with a coherent self-reference, can reason if they will be able to do some simple tasks etc.

Of course, you can be like Daniel Denett and bite the definition bullet - he was talking about 'free-will', not consciounsess and that a chess program has the necessary properties.

But, it makes more sense to take conscious experience as more fundamental, what we are directly aware of, and try to explain everything else with that as the base.