Philosophy, health, business, politics, economics, law, music — no matter what field you are planning on entering, artificial intelligence is sure to have an effect on it. Yet, university-level AI courses are typically restricted to technical audiences, making the topic inaccessible for much of the student body. For example, at the University of Michigan, you must take at least four computer science courses before you can take “Introduction to Artificial Intelligence” or “Introduction to Machine Learning.”
This barrier to entry, among others, is keeping out the diversity of perspectives that AI desperately needs. Already we are seeing racially-biased facial recognition systems, social media algorithms that encourage negative, primal content and increasingly autonomous weapons. We need leaders that can understand AI from a wide range of viewpoints — not the insulated system we have now. Relying on computer scientists to gain expertise in other fields neglects the value and expertise of non-technical experts and underestimates the complexity of these issues. Computer scientists are not often experts in anthropology, sociology, philosophy, health care, economics or law, but often there is an assumption that they can become experts in little time. This hubris is what has allowed Silicon Valley to threaten privacy, autonomy and democracy. We need experts from other fields who understand the basics of AI, not experts on AI who understand bits and pieces of other fields.
Additionally, many jobs now require a basic understanding of how AI works, what it’s good for and what its risks are. Without a good understanding, people may make poor decisions, rely on AI too much or too little and exacerbate existing biases — with major consequences. Just this week, a study found racial bias in a health care algorithm sold by Optum. The algorithm consistently underestimated the health needs of Black patients, potentially affecting millions of people. As AI enters more and more of our lives, we need people from a diverse range of perspectives who can reason about its impacts.
This is why I believe we need an AI course for a non-technical audience. I envision an introductory AI class, similar to introductory economics or statistics, that teaches the basics of AI to students from different majors. Creating such a class would acknowledge our changing world and prepare students for challenges they will face in their careers. Additionally, it would empower groups who are underrepresented in STEM, but whose perspectives are desperately needed, to access AI. Andrew Ng, a leading AI researcher and adjunct professor at Stanford University, recently created an online course called “AI for Everyone” that aims to teach non-technical people the basics of AI. More than 180,000 students have enrolled in it so far, demonstrating the appetite for the topic. At the University, a similar class would be a good match for our interdisciplinary academic environment and could be a model for other schools. The class could cover the technical fundamentals of AI, ethical and societal considerations, case studies involving its successes and failures and the future of AI. This would enable students to pursue AI further, whether in research or practice, gain skills and understanding that will be vital to their careers and contribute to important discussions about AI.
Also, by limiting comprehension to a small, relatively exclusive group of people, we are missing out on potentially transformational applications of AI. Opening up AI could lead to a flowering of applications in finance and medicine among other multifaceted uses that would not be possible if we relied on computer scientists alone. For example, an art collective created a piece with AI that sold for over $400,000 at auction. The University of Southern California Center for Artificial Intelligence in Society is using AI in areas such as homelessness and suicide prevention. AI is also assisting in the fight against climate change. Who knows what could be dreamt up by a generation of students with a new, immensely powerful tool?
We need an AI course for people of all backgrounds because the impacts of AI extend far beyond computer science. We have already seen the consequences of keeping this field closed off, and as the power of AI grows, so does the need for leaders who understand it. By making AI more accessible, we can mitigate its risks and enable innovative, beneficial applications, helping to ensure a better future for us all.
Chand Rajendra-Nicolucci can be reached at chandrn@umich.edu.