Limitations of Artificial Intelligence

Navigant Managing Director Jim Vint and Professor Gary Marchant of Arizona State University discuss the limitations of AI

Navigant Managing Director Jim Vint and Professor Gary Marchant, Arizona State University, discuss the limitations posed by artificial intelligence (AI) and how they will be addressed as the technology develops. To exemplify, AI machines replicate human functions, but they do not possess the same innate common sense that a human would. As a result, one risk of using AI is that the machine will simply miss something that is obvious to a human because the machine does not have any basis to know it, demonstrating a danger if AI is followed blindly. Other potential problematic features of AI include how the machines are accessing data, and if the data it is accurate, biased, proprietary, or privileged. Early adopters of AI must constantly ask if the machines are missing something potentially troublesome or if the data used has some implicit bias affecting the output. 

About the Experts

Back to top