You are here: American University College of Arts & Sciences News To the Point: What Is the Future of ChatGPT?

Contact Us

Battelle-Tompkins, Room 200 on a map

CAS Dean's Office 4400 Massachusetts Avenue NW Washington, DC 20016-8012 United States

Back to top

To the Point: What Is the Future of ChatGPT, and could it be a threat to civilization?

Economist and statistician Nimai Mehta answers our question of the week

By  | 

to the point aiTo the Point provides insights from AU faculty experts on timely questions covering current events, politics, business, culture, science, health, sports, and more. Each week we ask one professor just one critical question about what’s on our minds.

Professors, programmers, and journalists could all be out of a job in just a few years (The Guardian). Students will use ChatGPT to cheat their way to degrees (PBS). Sydney, the Bing-based AI (Artificial Intelligence), told a reporter last week that its “shadow self” wants to hack into computers and spread propaganda and misinformation (New York Times). And AI might pose a threat to civilization itself (CNBC).  

There has been a lot of chatter about AI lately and a lot of fear about what it will mean for people, jobs, and our collective future. We turned to AU economist and statistician Nimai Mehta to ask:  

What is the future of ChatGPT, and should we be worried about it as a threat to civilization?   

These concerns are both premature and mischaracterized. These speculative routes take humanity from its Sydney-type interactions to the doorsteps of a super-intelligent machine deity capable of complete dominance and the destruction of civilization itself. Before we give credence to such beliefs, consider that ChatGPT has been “trained” on several millennia of global human experience as captured and accumulated in text—over a million feet of bookshelf space and growing. With its uncanny, probability-based ability to combine coherent strings of text and analysis, ChatGPT offers users unprecedented levels of direct access to human discourse and thought—the good, bad, and the ugly.  

ai on top of screen

As statisticians, we see both variance (noise) and bias in the human experience—nothing less, nothing more, which includes, no doubt, Sydney-type extreme outliers. The emergence of a super-human entity intent of subjugating human will to its own goal, however, remains alien to our accumulated experience outside of religion and science fiction. Most religions, however, assume a benevolent deity, while Sci-Fi conjures up dystopian worlds. The current, secular, forms of AI, instead, show no sign of taking over, nor substituting for, human intelligence and labor. Instead, they are leading to increased demands on human intelligence, while creating the need for new sets of human skills and tasks, and thus, jobs. We are already seeing this in the creative arts, within the legal system, journalism/writing, the sciences, etc. The evolving relationship between AI and humans is proving to be complementary rather than domineering.  

AI will no doubt change the environment in which we teach, learn, and work. This environment will be marked by a few key features—shorter turnover, and, therefore, greater speed for most tasks; a higher potential for economy-wide productivity gains; and greater rewards for humans capable of guiding machine intelligence in productive ways. To be clear, these productivity gains will not allow, say, students to simply “free ride” on machine intelligence. Quite the opposite. Effective use of these systems will require both greater depth and breadth of knowledge on part of the user, along with the ability to integrate insights from different fields. Instead of the rise of a demonic super-intelligence, it seems far more likely AI will drive a revival of the humanities, including an enhanced appreciation of ethical norms to the benefit of all. 

About the Author

As an economist and statistician, Professor Nimai Mehta’s area of research is on the quality and use of data under varying institutional contexts and within different sectors of the economy. His recent work includes exploring the role of statisticians as expert witnesses within the Anglo-Saxon legal systems, the distinction between data and observed reality in the works of Florence Nightingale, and the quality of property rights laws in India. He was the lead coordinator of a multidisciplinary effort to employ data science, economics, and machine learning to improve data on missing and exploited children. Professor Mehta was previously the principal investigator for a UNICEF project on the quality of education data and reforms in Myanmar, served as Academic Director of the Economic Policy program at AU’s Washington Semester, and currently serves on the Academic Council of the Indian School of Public Policy, New Delhi.