Written by Robby Mangum
On April 10, Dr. Caleb Bradberry, assistant professor of Information Technology at Radford University, hosted an MSCP conversation about the ethics of machine learning and data collection. Dr. Bradberry, an assistant professor in the Information Technology Department at Radford University, discussed the applications of machine learning for predictive algorithms and the underlying philosophy behind this topic.
Machine learning is basically teaching a computer to ‘think,’ or be able to make decisions based on previous data and observation.
“If I walk forward (toward this table),” Dr. Bradberry said, “I’ll hit it. But how do you teach a machine to make a prediction like that?”
Dr. Bradberry listed off several applications for machine learning: predicting heart troubles in patients, predicting whether a student will drop out of college, or predicting stock prices.
“’The Philosophy of Science,’” he said, “is how we can inherently ‘know’ something to be true.”
He said that its application to machine learning is examining how, exactly, artificial intelligence can be represented. These machines have to exercise epistemological or ontological philosophy, according to Dr. Bradberry, in order to make decisions: determining how a thing is or what a thing is, respectively.
Finally, Bradberry discussed the ethical implications of machine learning. Many platforms such as Google and Facebook collect data from the people that use these platforms. While this is outlined in the terms of service, few people read that and readily consent. Dr. Bradberry also pointed out that machines can exhibit bias.
“A sample size of people from Silicon Valley is going to be predominately white and male, so a machine that learns from its surroundings may think that, to be a person, you must be white and male,” said Dr. Bradberry.