Artificial Intelligence is a nascent technology poised to transform the role of talent in the modern economy. From driverless cars to intelligent machines like IBM’s Watson, AI is likely to displace jobs and create new ones.
But with the new technology come new skills needed to use it, and just as companies like Google, Uber and others have hired AI experts to develop forthcoming products and services, colleges and universities have been quick to offer new courses on the subject.
So what skills do AI courses teach?
First, students learn the fundamentals of search. In computer science, search involves looking for a particular solution to a given problem, according to Daniel Tauritz, associate professor and associate chair for undergraduate studies and outreach activities for the department of computer science at Missouri University of Science and Technology at Rolla.
In his Introduction to Artificial Intelligence course of about 50 students, Tauritz covers the following three basics of search:
Uninformed Search: This is used when creating an action sequence that doesn’t account for any changes along the way.
Heuristic Functions: These allow for decisions to be made without accurate or complete information.
Adversarial or Moving Agent Search: This is used when there are other entities making decisions that influence one another.
Piotr Gmytrasiewicz, associate professor in the department of computer science at the University of Illinois at Chicago, teaches three courses: Artificial Intelligence 1, Artificial Intelligence 2 and Applied Artificial Intelligence.
Artificial Intelligence 1 covers logic-based approaches, while Artificial Intelligence 2 showcases numerical and mathematically focused approaches based on probability theory.
Applied Artificial Intelligence, which contains about 30 students, concentrates on at least five applied AI technologies that are most prevalent in the technology industry today:
- Rule-based system: a logic-based approach.
- Probabilistic system: knowledge using probabilities.
- Fuzzy Logic system: logic that’s neither true nor false.
- Decision-making system: computes the best decision in a problem.
- Machine learning system: A neural network or a decision tree which can be learned from data.
A Changing Field
Throughout his roughly 20-year teaching career, Gmytrasiewicz said college courses covered logic and search techniques when teaching AI. Now, however, there are many more quantitative and mathematical approaches, as probabilistic approaches drive the current applications that dominate the AI industry.
Tauritz said one newer important skill area is ethics, which is important in AI because humans are programming intelligent machines that could make potentially life-altering decisions. With autonomous cars, if there’s an accident, the machine might need to make decisions that place value on lives. The trolley problem is a common example of this type of decision-making: If there’s a trolley speeding toward you, and its brakes are out, the multiple passengers on the vehicle might die, but you can switch the track so they reach safety. However, there’s man on those tracks you’ve switched the trolley to, and he’s about to get run over. The passengers will be safe, but the man will die. This problem can take other forms that require decisions on other lives, which makes it a teaching tool on the philosophical and ethical issues that arise.
In some instances, AI technology is doing the teaching on the subject. The Knowledge-Based Artificial Intelligence course at the Georgia Institute of Technology uses its own “Jill Watson,” who helps respond to repetitive questions on the course’s online discussion forum. Graduate students and Ashok Goel, a College of Computing professor at Georgia Tech, created the AI partly using technologies from IBM’s Watson platform, Georgia Tech’s News Center wrote.
Other skills and qualifications listed on job postings for AI-related roles include:
- “A strong passion for theoretical and empirical research” for a research scientist role at Facebook
- “Fiction and non-fiction writing” for a cognitive scientist/engineer position at Raytheon
- “Exposure to research in neural nets such as memory augmentation, reinforcement learning and representational learning” for an AI computer scientist job at General Motors
- “Direct contribution to the design and implementation of innovative products that were successfully shipped and used by consumers” for a machine learning engineer, proactive intelligence role at Apple
- “Instinctive knowledge of how people will want to use a product, service, or solution that solves a real world business problem” for an applied machine learning software engineer expert at Allstate
- “Experience quantitatively evaluating algorithm performance, demonstrated via peer-reviewed journal or conference publications, or technical reports” for an autonomous driving research scientist position at Ford Motor Co.
Lauren Dixon is an associate editor at Talent Economy.