Artificial intelligence is now used in virtually all aspects of our lives. Yet unchecked biases within existing algorithmic systems, especially those used in sensitive use cases like financial services, hiring, policing, and housing, have worsened existing societal biases, resulting in the continued systemic discrimination of historically marginalized groups. As banks increase AI usage in loan and appraisal decisions, these populations are subjected to an even greater precision in denials, eroding protections provided by civil rights laws in housing. Meanwhile, the use of facial recognition technologies among law enforcement has resulted in the wrongful arrests of innocent men and women of color through poor data quality and misidentification. These online biases are intrinsically connected to the historical legacies that predate existing and emerging technologies and stand to challenge the policies created to protect historically disadvantaged populations.