If you want to justify your position, I would suggest highlighting the positive changes you _did_ make, rather than the positive changes you _intended_ to make. If you didn't make any notable positive changes over the time you worked there, that might explain why you no longer do work there.
'Ethical AI' in my opinion is essentially the AI version of 'positive discrimination', i.e. they will keep biasing the inputs and outputs until the algorithm yields the desired result. In this case the desired result in not equality, but "corrected equality". Don't get me wrong, biased and misleading training sets are a problem, but the AI itself is not unethical.
Also I would like to know whether algorithm transparency included blue check-marks (that were obviously biased), or government take-down requests, or the promotion of completely random people for no apparent reason? There was never any intention for algorithm transparency, this is Twitter's 'secret sauce' after all.