It will definitely have some errors, but I scanned the results for the first 100 movies before I started importing them, and I think the value-add will be much greater than the number of errors.
Does Wikidata have a quality goal or error rate threshold? For example, Freebase has a nominal quality goal of 99% accuracy and this is the metric that new data loads are judged against (they also want to be in the 95% confidence interval, which determines how big a sample you need when doing evaluations).
I haven't looked at this bot, but a develop/test/deploy cycle measured in hours seems, on the surface, to be very aggressive.