The fast pace of change in AI technology is such that we cannot wait any longer to address concerns linked to a series of ethical questions, including the grouping and interpretation of data by machines which may lead to undesired and biased results and discrimination, for example in assessing for health insurance or profiling a subject in the area of criminal justice.

Also, according to many, including the United Nations University (UNU), Ai is tied not only to algorithmic design, but also to the datasets used to train it: When embodied in ubiquitous facial recognition systems, criminal justice predictions, job selection, or in the functioning of hospitals and financial markets, data quality and integrity risks are amplified. One example of this challenge is the use of facial recognition and predictive policing algorithms by law enforcement when trained on historical crime data, platforms like PredPol and Palantir create negative feedback loops that recommend disproportionate surveillance in low-income and Black communities.

Briefly research some example(s) of AI biases around gender or race. What was at the root of the bias? Do you think the AI decisions and actions resulting in bias was the result of embedding ethical values into algorithms by the humans and therefore they should be held accountable? Or do you think Goodwill alone will not suffice and there should also be government regulations or enforced standards and industry wide governance making sure unconscious biases are removed in the evaluation of data and ensuring transparency in the way every AI system reach a conclusion.