Yes, with the nuance that companies can put differing levels of effort into trying to avoid the generation and/or delivery of results that reflect unpleasant th

Yes, with the nuance that companies can put differing levels of effort into trying to avoid the generation and/or delivery of results that reflect unpleasant things in the training set. AI is just data+math; it can't be any more or less biased than the data and math. But the companies have a lot of knobs to turn to either mitigate that or not.