Advertisement

Colleges must prepare students for an AI world

JPL engineers put together a drone race to find which is faster – a drone operated by a human or one operated by artificial intelligence. The race capped two years of research into drone autonomy funded by Google.

Computer scientist Andrew Ng has called artificial intelligence “the new electricity” and bemoaned a “scarcity of talent and data slowing” its adoption. Meanwhile, research and advisory firm Gartner has projected that AI will create 2.3 million new jobs next year — while eliminating 1.8 million jobs.

Our college curricula must respond swiftly. As educators, we must prepare our students to succeed in a society where human and machine intelligence co-exist. Corporate human resource executives also must consider how to train employees to be productive in the new workplace.

Advertisement

The rise of artificial intelligence and machine learning (AI/ML) creates new imperatives for what we teach across all majors. Much has been written about how AI technologies can be leveraged to deliver education, for example, by helping create material that is more personalized, flexible and responsive.

Montreal’s Dawson College recently budgeted $1 million for an AI initiative that will engage some faculty — the project’s first fellows — in AI-related integration that will transform the education of all students.

New technologies are clearly relevant to students enrolled in computer science, engineering and other technical disciplines. But a move like Dawson’s signals the impact extends far further.

Indeed, three critical elements must be part of any university education today.

First, every student should be required to take a course on AI/ML — call it AI Literacy 101. The course will demystify AI, separating hype from reality. It will discuss the original, grand vision of AI; the missteps along the way and how we got to where we are now.

Students will learn about various technologies that constitute AI — robotics, computer vision, natural language processing and more. They’ll discuss the Turing test: the touchstone that allows us to determine whether machine intelligence is distinguishable from human intelligence.

They’ll explore the limits of AI — what it can and cannot do, where it has the potential to be useful, and what aspects of work and decision making remain within the exclusive purview of humans. Doubtless, such information will be pragmatically useful in guiding future career choices of fields of study.

Second, while the benefits of diversity and ability to work in diverse teams have always been important, they become front and center in an AI/ML world. That’s because AI/ML relies on vast quantities of data to learn — to recognize patterns, make predictions and understand the relationships among factors.

What the machine knows is what it is exposed to. Just as humans who interact with only those who are similar to them will be less likely to develop a broad knowledge base that equips them to address unanticipated scenarios, likewise, a machine algorithm that gets data on only one type of individual or situation becomes less likely to perform optimally when encountering an unexpected situation. A wide-ranging set of experiences makes us smarter as humans and, correspondingly, diverse data make algorithms better.

Being an informed consumer of AI/ML requires every student to have a deep understanding of the nature and existence of bias and how it can be minimized or eliminated in these new intelligences we create.

Third, we must prepare students to work in concert with machine intelligence — whether in a substitution or supplemental mode.

Should AI replace a human? In robotics, the choice is simple — if the robot is capable of emulating human motion, cognition and response to an acceptable level of performance, it will.

The question is more nuanced in work that involves judgment and intuition. AI may be used to complement human analysis. In radiology, for example, algorithms today detect anomalies and human radiologists confirm them.

Future workers must have keen appreciation for where the human-machine boundary falls. Just as it is a learned skill to work effectively in teams with smart colleagues, working hand-in-glove with AI will need to be taught.

Advertisement

Change is needed now. Progress in these technologies will continue to unfold exponentially. We simply must accelerate our response in higher education.

Ritu Agarwal (ragarwal@rhsmith.umd.edu ) is a professor and senior associate dean for faculty and research at the University of Maryland’s Robert H. Smith School of Business. She is also the Robert H. Smith dean’s chair of information systems and will serve as the school’s interim dean, effective Aug. 15.

Advertisement
Advertisement