AI and the Great Chessboard of Humanity

When IBM's Deep Blue supercomputer defeated reigning world chess champion Garry Kasparov in 1997, some said it was the beginning of a long-predicted takeover of the human world by Artificial Intelligence. Similar concerns were expressed when the internet was invented, as well as the automobile, the telephone, electricity, and the printing press. None of these inventions had the catastrophic effect on humanity that some feared they might. Yet the fear of technology rendering humans useless persists, even as we continually seek to ease the work of existence by automating as many tasks as possible.

ChatGPT, an AI chatbot programed to respond to questions in natural human language text, stunned the world with its elegant communication when it was released in November 2022, gaining more than 150 million users worldwide in the first six months. GPT (Generative Pre-trained Transformer) is a large language model trained on vast amounts of text data to learn linguistic patterns and context. Transformer architecture, based on an early understanding of neural networks, enables GPT to perform these tasks and generate text that closely resembles human conversation.

Take it out for a test drive on any subject, and the results are impressive. For example, ask ChatGPT why humans continue to play chess, since computers have already solved the game. It responds with a list of thought-provoking points followed by a summary that reads, "[D]espite computers' ability to solve chess, human engagement with the game persists due to the enjoyment, social aspects, learning opportunities, creativity, educational benefits, and cultural significance it offers. Chess continues to be a vibrant and captivating pursuit for players of all levels

This is remarkable, considering the program uses algorithms and statistics to predict the next word in a sequence. Processing one equation at a time, at lightning speed, it formulates a humanesque response in seconds flat.

This conversation begs a follow-up question: do computers enjoy playing chess? ChatGPT's reply offers important insights about the potential and limits of AI, edited here for brevity: Computers are not capable of experiencing emotions or subjective enjoyment as humans do. They do not possess consciousness or the ability to have preferences, desires, or feelings.

Computers are programmed to execute tasks efficiently and effectively, including playing chess. They are designed to analyze positions, calculate moves, and make decisions based on algorithms and evaluations. The "enjoyment"; for computers lies in successfully executing their programmed tasks and achieving the goals set by their designers.

They excel at finding the best moves based on precise calculations and evaluations, and their ability to process vast amounts of data quickly. While computers may not experience enjoyment in the way humans do, they can provide formidable opponents and offer challenging gameplay experiences for human players.

Notice the subtle superhuman braggadocio in ChatGPT's description of computers as "formidable opponents." To be clear, this phrase is the output of a series of data inputs and mathematical equations that predicted its likelihood of being the correct phrase given the context. It is this very close AI approximation of human thoughts and feelings embedded in the words humans choose for self-expression, made possible through mathematical programming, that has prompted some of the original pioneers, architects, and most ardent supporters of AI to plead for caution and restraint as the tech world dives deeper into its exploration of artificial intelligence.

However, it's important to note that this same ChatGPT response highlights the limits of AI as incapable of human emotion, desire, and consciousness, which are uniquely human abilities that drive creativity and innovation. In The Fallacy of Obviousness (Aeon, 2018), Dr. Teppo Felin, Douglas D. Anderson Endowed Professor of Strategy & Entrepreneurship at the Huntsman School, explains that distinctively human characteristics are essential in making sense of the world around us, and that computation alone does not "exhaust or fully capture the human mind." "[I]ntelligence and rationality are more than just calculation or computation and have more to do with […] deciding what is relevant and meaningful, and what is not…And relevance and meaning continue to be outside the realm of AI."

AI is here to stay, and ChatGPT, programmed to execute tasks and achieve goals set by designers, is one of many AI-driven tools that have already entered the workplace. But human ingenuity determines the next move. Understanding the capacity of AI and specific tools like ChatGPT to checkmate or champion humankind and determining the strategic moves to advance their superhuman potential in our favor, are pressing matters for higher education.

Ally or Opponent

The challenges and opportunities of AI in higher ed are interconnected around the fundamental concerns of labor displacement, ethics, and the evolving landscape of teaching and learning in an AI-driven world.

Labor Displacement

AI is already automating mundane tasks and redundant processes. Certain tasks traditionally performed by humans will be overtaken completely by AI. But rather than cause job loss en masse, this will redefine the nature of daily work. Increased automation will ease administrative tasks, increase operational efficiency, maximize human productivity, and make room for more creative and meaningful work. In turn, this will increase job retention and satisfaction. It will also improve the student experience and educational outcomes as professors, administrators and staff are more available to help solve problems and provide guidance.

Dr. Ben Blau, George S. Eccles Endowed Chair in Finance and Department Head of Economics and Finance at the Huntsman School, is optimistic about the impact of AI on the labor market. "AI can help ease scarcity of labor as birth rates fall. And, as with agriculture and manufacturing, history demonstrates that we continue to innovate after repeated disruptions. Rather than harm the labor market, AI will more likely create a net gain in job opportunities."

Ethics

AI is streamlining the admissions process by sorting applications and helping recruit students to matching schools. Chatbots improve academic success, retention, and the overall education experience by answering questions about everything from financial aid to course descriptions, and sending reminders about class schedules, upcoming tests and deadlines. But these processes also raise ethical concerns about privacy, data security, bias, and fairness in the algorithms that determine the type and amount of personal data collected, and about the ways in which this data might be used to impact educational outcomes.

The ease of plagiarism with ChatGPT is particularly concerning, as this technology decreases the need for students to develop critical thinking skills and, when relied on too heavily, deprives them of the joy of discovery that is fundamental to learning.

Additionally, the data on which ChatGPT was trained does have its limitations because it only tells part of our collective story. Dr. Sharad Jones, Professional Practice Assistant Professor in the Department of Data Analytics and Information Systems at the Huntsman School, explains, "ChatGPT is built on the Transformer, which began as an open-source development which allowed everyone to actively study the capabilities, risks and harms of AI. The model was trained using a massive amount of computing resources and scraped internet text data, which tends to reflect a predominantly western, industrialized, and educated perspective. Given the scale of the data, much of the content moderation prior to training was performed automatically, but this process raises concerns about the exclusion of certain cultural perspectives and the risk of a homogenized user experience. With the release of the latest version of the model that backs ChatGPT (GPT-4), specific details about its architecture and training data were kept private due to 'the competitive landscape and the safety implications,' only furthering the difficulty of understanding the potential impacts of these tools."

These ethical considerations do not mean that we should rush to ban ChatGPT from higher ed. Rather, we can feel confident using ChatGPT as a powerful tool to help with a variety of administrative and academic tasks, and at the same time remain skeptical and inquisitive about the text it produces.

Teaching and Learning in an AI-Driven World

Maximizing the positive potential of AI in teaching and learning will require thoughtful efforts to balance technology and human interaction.

"AI will democratize education, which will have a huge and positive impact on education worldwide," explains Dr. Pedram Jahangiry, Professional Practice Assistant Professor in the Department of Data Analytics and Information Systems at the Huntsman School. "Underprivileged students will have increased access to knowledge, and tools like ChatGPT can provide a personalized learning experience for students at all socioeconomic levels by acting as a private tutor with perfect patience and the ability to adapt to students' needs and level of understanding in real time."

While AI-enabled education tools offer real value, Dr. Chris Corcoran, David B. Haight Endowed Professor of Analytics and Department Head of Data Analytics and Information Systems at the Huntsman School, is cautious about removing the human component. "Many online learners lack the resilience and discipline to commit to self-guided learning. Memory is three dimensional, and retention is lower when reduced to two dimensions. Furthermore, learning is a social endeavor. Humans are not wired to be isolated. We need deadlines to help us stay on schedule, and a community of learners to bond together and model behaviors for each other. Students also need mentors and connection to the broader community for professional development and success."

At the heart of this issue is whether individuals choose to engage in the transfer of information or in the production of knowledge instead. Information transfer can be measured in exams and essays. Individuals who excel at rote-memorization have historically risen to the top in this model, but AI can transfer a greater quantity of information and at a faster speed than a human can transfer information. By contrast, the uniquely human characteristics of curiosity and creativity will drive the production of knowledge in an AI-centered world.

Strategic Moves

ChatGPT's self-described abilities and limitations above provide insight into what humans will do in an AI-enabled world and how best to prepare them to do it.

AI can "execute tasks efficiently and effectively; make decisions based on algorithms and evaluations; and execute their programmed tasks and achieve the goals set by their designers." AI cannot "possess consciousness," and must obey the rules of human designers. In other words, AI in its current form depends on human intelligence.

"[T]he role of human ingenuity in crafting questions, expectations, hypotheses, and theories to make sense of their environments…[is the] approach that gives humanity its due," writes Felin. "The human ability to ask new questions, to generate hypotheses, and to identify and find novelty is unique and not programmable. No statistical procedure allows one to somehow see a mundane, taken-for-granted observation in a radically different and new way. That's where humans come in."

Rigorously Pursue Human and Artificial Intelligence

Experts in the Huntsman School of Business agree that the first step is teaching students how to use AI tools. Many professors in the School have already incorporated ChatGPT and other AI into their curriculum.

"Goldman Sachs is now hiring people for a role called Prompt Engineer. They want people who know how to talk to GPT, and that's exactly what I'm trying to do in my classroom—teach my students how to coexist with AI and communicate with it," explains Jahangiry. Prior to ChatGPT, he taught students how to code as part of his machine learning classes, but now focuses more heavily on the theory of deep learning and helping his students learn to communicate with ChatGPT to produce code and study problems. "Every single person with average talent can now code decently with GPT, so students need to switch focus. We need to reshape the methodology of our teaching to help students develop the tools to work with AI, to know what to ask, and to know how to communicate effectively with it to get results."

"Our students are going to use GPT for the rest of their lives. They need to learn how to do it here, in an environment where we have professionals and a community around them to help them learn to do it in the best way," explains Dr. Andrew Brim, Professional Practice Assistant Professor in the Department of Data Analytics and Information Systems at the Huntsman School. "If they want to be competitive in the marketplace, they're going to have to be able to use this tool. It's not going to do the job for them, but it's a tool that will help them get to the rest of the job."

Critical thinking—the ability to analyze, evaluate, and make judgements—is a defining characteristic of human intelligence that must be constantly reinforced throughout the educational experience. Corcoran explains, "Our department is doing a lot of self-examination and asking, what are the learning objectives of this class? What do we really want students to be able to do when they leave this class? What do they need to understand to be qualified for the world of work and make a contribution? The solution is going to be about understanding the process and thinking critically."

Jones uses ChatGPT in the classroom together with his students to develop ideas and sharpen research questions. He strives to help students develop the intellectual curiosity to ask deeper questions that broaden thinking and encourage serendipitous moments of discovery. "Every time GPT returns an answer, I encourage them to ask Why?, and then embrace the random moments of connections. Let's see where this leads."

Brim also uses computers to help students develop critical thinking skills. Together, they examine logs containing previous iterations of a student's work, looking for common trends and gaps in questioning to help the learner better understand the thought processes that led to discovery or dead ends.

These conversations about thought processes are also more effective as evaluation tools than exams or traditional homework assignments. Jones explains, "Rather than education as solely the transfer of facts to be consumed, applied and extended, the education platform of the future will focus on the process of learning. If we can put the traditional concept of 'grading' aside for a moment and focus on our goal, much of the value that we can provide is helping students develop a learning process."

Develop Uniquely Human Skills

"Our department has a very healthy perspective about AI, which is built on our belief that no matter where it takes us, we're still going to have to work as a team. We have to be able to communicate effectively with other people. We have to learn to collaborate so that we can solve problems together," says Corcoran.

In connection with the philosophy department, the Huntsman School will roll out a course to address the intersection of AI and ethics. "Much of the conversation about AI is positioned within a reactionary framework about how AI is causing, or will cause, harm. This framework is not productive, and doesn't produce solutions," explains Jones. "This new course will give students the vocabulary of philosophy and ethics to be able to talk about these issues at a fundamental level, and then examine case studies outside of that reactionary framework to explore the real opportunities AI technology creates."

Brim also begins each new class with a broader discussion of the nature and purpose of intelligence. "The purpose of human intelligence and human life should be our roadmap for AI. If the purpose of human intelligence is to build community, solve problems, and improve lives, then artificial intelligence should run parallel to that purpose."

When "thank you" was typed into the query bar at the conclusion of the ChatGPT conversation about computers and chess, it responded with, "You're welcome! I'm glad I could help. If you have any more questions, feel free to ask." The exclamation mark denoting energy and enthusiasm is as startling as a computer referring to itself in first person and expressing a willingness to be helpful, which are human characteristics. This exchange helps us better understand the value of developing soft skills.

Brim reminds us, "AI may say 'I love you,' but it's a lie. It doesn't possess consciousness, and it's incapable of emotion. Only a human has the ability to put you above themselves, and to empathize. AI can say the words, but it can never replace real human interaction." In a world where AI can manage redundant processes, the soft skills belonging to human consciousness and exhibited in real-time human interactions will become more valuable. Individuals who possess empathy, confidence, and flexibility, and who excel at communication, teamwork, time management, networking, attention to detail, problem solving and conflict resolution will complement the strengths of AI and position themselves to lead in an AI-enabled world.

Experiential Learning

Professional practice faculty in the Huntsman School possess the industry experience to provide students with the latest techniques and translational ability to apply innovative ideas to problems in an organization. Before joining the Huntsman School, Brim worked in software development and derivative trading for ten years, and Jahangiry was a researcher in financial modeling at BlackRock. Jones worked as a data scientist in space technology and also helped design Carvana. Such real-world experience helps them mentor students and provide them with relevant knowledge and skills.

The Analytics Solutions Center, housed in the Department of Data Analytics and Information Systems, gives students the opportunity to work together with faculty on projects for real companies. In one project, students were tasked with figuring out how to detect when a spherical fireball began to lose shape, to ensure both the physical safety of the company and free employees from the tedious task of keeping watch in order to engage in more fulfilling work. Together with Jones as their project mentor, the students developed the tools to solve the problem and delivered a solid solution to the client. Because of this hands-on experience, one of the students landed a job as a machine learning engineer.

"The Analytics Solutions Center gives students exposure to companies, and real-world experience in solving problems. The companies get to have the real problems actually solved, and they're able to see student execution and maybe hire them directly. It's a win for both sides," says Brim.

The essence of experiential learning is helping students learn how to learn. It's not possible to know all of the tools or methods to solve a problem. The Analytics Solutions Center allows mentors to model—together with students—the methods to arrive at a solution. Jones explains, "My goal is to empower students to feel confident that they can learn anything. I tell them to use every resource at their disposal, and then help them discover the process of solving the problem."

Corcoran agrees. "Our goal is to help students have confidence. The way to do that is to be here in a community, facing these things together and sharing different perspectives."

The Final Say

Adam Smith famously argued that "in the great chessboard of human society, every single piece has a principle of motion of its own." We are independent creatures of consciousness and ingenuity, with abilities, interests, and experiences uniquely our own. We are interconnected by a complex interplay of unpredictable forces, and we share in common the ability to comprehend and adapt to change. It is these features of humanity that will enable us to harness the power of AI and master the game of life.