Suppose a company is considering developing and implementing an algorithm to automate a process and reduce costs. Given the evidence on algorithmic bias and discrimination, point out potential pitfalls and what the company can do to try to avoid them. Furthermore, discuss what solutions may work better or worse in different contexts.