I think you need to relax the idea of 'detecting life' towards 'detecting evidence of life', i.e. evidence of populations of locally-entropy-defeating systems that reproduce.

Once you relax the requirement that 'answers' must be 'yes or 'no' to 'more likely' or 'less likely', it should be clearer how one would go about searching for life. For one, is there chemical activity one wouldn't otherwise expect if there was no life?