As a senior fellow at UPCEA, the online and professional education association, and a professor emeritus at the University of Illinois Springfield, I am privileged to contribute to the latest efforts to track, analyze and present the impact of technology on higher education. From my first appointment in 1972 and continuing over the next 52 years, I closely followed, managed, researched, published and taught about the use of technology to improve learning and teaching in higher education.
It’s been an exciting journey from analog technology like film and tape sound of the early years to the digital transformation that brought the web and online learning to students and faculty around the world supported by a multitude of platforms and supporting technologies. Now, at the convergence of productive AI technologies that bring agency, autonomy and robotic uniformity, we are facing a moment that promises to eclipse that half-century of development and reform in higher education in the space of the coming months and grow over the coming years.
To effectively envision the near future of generative AI in higher education, we need to at least briefly consider the broader economic context, the general funding situation for higher education, the state of development/deployment of new and emerging AI technologies, and the emerging demand for graduate students, highly skilled and re-skilled workers with certifications from colleges and universities.
Overall, we are seeing evidence in the wider world of business and industry that there is interest in AI as a way to achieve greater efficiency and effectiveness, even in roles traditionally held by humans. Change is already happening in the technology sector. In this column, we’ve documented cases where leaders of companies like IBM, Cisco, Microsoft and TurboTax have justified massive layoffs to introduce AI programs and have AI take over jobs previously held by humans. In the article, “Generative AI Update for 2024,” in European Business Review earlier this year, my colleague Katherine Kerpan of the University of Illinois Chicago and I documented the beginnings of this movement, including behavioral support strategies for workers with outdated, inefficient skills and career paths. Suffice it to say that the competitive forces of efficiency and innovation are driving the adoption of these technologies beyond the academy.
Meanwhile, a large number of higher education institutions suffer from low income and operational constraints. Last month, Forbes released its “Forbes College Financial Grades” list, noting, “Nearly 55% of schools, or more than 480, received a C or worse, compared to only 20% in the 2021 fiscal year. One hundred and eighty-two schools received a D, the lowest Grade, out of 20 in 2021. Earlier this year, John Marcus wrote in a book The Hechinger report that “Colleges are now closing at a rate of one per week.” Marcus writes that in the vast majority of cases, surprising students are left with a difficult road to graduation and certification. Student loan debt continues to be staggering, currently at one and three-quarters of a billion dollars! The number of emerging demographics or student enrollments scheduled to reach higher education in the following year. Because of these factors, there is some awareness and concern in our field that we must work more smoothly and effectively to meet the expectations of potential students and the increasing competitive pressure in our field as the number of institutions decreases.
Research and development across the broad field of artificial intelligence takes place at thousands of institutions and startups around the world. The recent release of OpenAI o1 just happens, as I write this, in a long list of growing developments across platforms from some of the world’s biggest technology companies to unleash the power of AI in a variety of ways and means. . If we take just one new development, we see the advent of level-two thinking. In a report accompanying the release, OpenAI writes, “OpenAI o1 ranked in the 89th percentile of competitive programming questions (Codeforces), ranked among the top 500 students in the US in the USA Math Olympiad (AIME) qualifying competition, and outperformed humans .PhD-level accuracy in benchmark problems in physics, biology, and chemistry (GPQA).
The report continues, “We also tested the o1 on GPQA diamond, a rigorous forensic benchmark that tests technology in chemistry, physics and biology. To compare models with people, we hired experts with PhDs to answer GPQA-diamond questions. We found that the o1 outperformed those human experts, becoming the first model to do so in this benchmark. These results do not mean that o1 is more powerful than PhD in every way—only that the model is more adept at solving certain problems that PhD is expected to solve.”
While o1 thinking is on the rise, we’re seeing the rise of artificial intelligence agents that aren’t simple chat bots. Instead, the agents that will flood the market this fall and beyond are capable of performing complex, multi-step, dynamic tasks. As Amazon Web Services explains,
“An artificial intelligence (AI) agent is a software program that can interact with its environment, collect data, and use the data to perform tasks for itself to meet predetermined goals. Humans set goals, but the AI agent independently chooses the best actions it needs to take to achieve those goals. For example, consider an AI agent for a contact center you want to solve [sic] customer inquiries. The agent will automatically ask the customer different questions, look up information in internal documents, and respond with a solution. Based on the customer’s responses, it decides whether to resolve the question itself or refer it to someone.”
More complex tasks can also be accomplished. We have seen many experiments using such agents in Minecraft as described in Tom’s Guide. Many communities have been built and interesting communities are built by intelligent agents given goals and objectives by humans. They organize and in some cases exercise democracy.
That leads us to look at higher education next year. Given this background, join me in envisioning how we can begin to use this technology. I see us replacing middle-level managers with intelligent agents who can make well-documented decisions that adapt to changing goals and outcomes. Areas such as admissions, financial aid, division of financial affairs, resource planning, human resources and many more are offices where some employees may first become artificial employees.
As scary as it may seem to some, I see these advanced models, like those with Ph.D. thinking, filling adjunct faculty positions while being supervised by human professors. The long-term Khanmigo project funded by OpenAI shows that valuable teaching, learning and personalization capabilities can be delivered through generative AI.
In some developing campuses, I see intelligent robotic agents by the end of 2025. I envision intelligent robots working side by side with students, faculty and administrators in the library, dining halls, health services, international student services, the physical plant. , campus grounds and many other units.
I hope you will follow the link excerpts to learn more about the topics in this column. Then, perhaps, you will start to form your own idea of how and when this technology will start in your university. This vision will help you inform the future of your university and your personal career plans.
Source link
