Should developing technologies that discriminate against racial minorities be an ethics violation?

While ethical engineers don't intentionally develop technologies to discriminate against racial minorities, some of the technologies they develop do discriminate.  Pulse oximeters don't work as well on patients of color.  Bias is a core problem of many AI applications. 
  • Should developing technologies that discriminate against racial minorities be some type of ethics violation? 
  • In what circumstances? 
  • Should engineers have an ethical responsibility to diminish any bias?