AIT-04 - (Unmasking Subjectivity: The Complex World of AI in Pain Assessments) Round II: The Ugly, Maybe Even COMPUTERS Have Race, Gender, and Class Bias

For the past 20 years, we have been obsessed over what the world will look like when humans are replaced by computers. Look at our love of movies like Terminator, 2001: A Space Odyssey, Blade Runner, and the Matrix. While most of these movies predict terrifying consequences, we have begun to integrate artificial intelligence into our medical practices: email programs and electronic medical records finish our sentences, our phones and social media have AI bots which answer medical questions, our phone trees are even being automated. As we begin to integrate these technologies into acute and chronic pain care, there are many hopes. One is that it will reduce biases and ensure that care is more consistent across different races, genders, social classes, and sexual orientations. While these seems likely, are we facing another Skynet?

Does artificial intelligence propagate or eliminate human biases?

In this course, we will review the theories and creation of artificial intelligence. We will also discuss technologies that read language, scan facial expressions, and predict illness. We dive into whether bias is actually eliminated when we allow computers to replace our human compassion. We will challenge the very premises behind the implementation of machine learning and offer potential suggestions for implementing it into your practice ethically. Lastly, we will answer the question: “Will artificial intelligence merely propagate disparities that already exist?”