AI in College: A Tool Without Training
Pam Sornson, JD
May 7, 2024
Artificial Intelligence (AI) is permeating every corner of society and has been for a few years. That reality is driving demand for AI technicians and programmers in almost all industries. Unfortunately, while the use of AI is accelerating, the development of a well-trained AI technical workforce is lagging far behind, and companies are clamoring for workers who can help them achieve their AI goals. In many regions, including California, the local community colleges are often seen as ground zero for the generation of new educational paths to AI comprehension and mastery.
‘Use’ Does Not Equal ‘Understanding’
For many AI users, early engagement with the advanced technology didn’t even register as an event. As organizations embraced and embedded into their services the cost savings of chatbots for customer service and robots for manufacturing, their consuming public was left out of that adoption information loop. Instead, those customers and colleagues simply continued to engage with the entity, unaware that they were now working with a digital program and equally unaware of the technological advances being made behind the scenes.
That reality reveals the challenge that many enterprises are dealing with today: using the technology is – at least for the moment – infinitely easier than introducing it to existing systems and workflows. AI’s architecture is different from traditional computer programs and apps and, in many cases, renders those legacy activities obsolete. Consequently, many company owners, agencies, and organizations are seeking resources to assist them in designing and implementing new AI systems into their existing infrastructures and then maintaining those capacities over time. In short, they are looking for an AI-savvy workforce that does not yet exist.
New Programs Require New Practices
America’s higher education system also faces a shortage of AI instruction courses. While many schools currently use the technology in at least some of their departments and divisions, there are very few training opportunities available to ensure new users can master and optimize their use of the latest digital tools. One researcher clarified the extent to which both K-12 and higher education schools use AI now and provide training for it:
Only one in four survey respondents acknowledge that their school has put intentional limits on the use of AI (specifically ‘generative’ AI) by its educators and staff. Eighteen percent (18%) report that their school is working to identify appropriate use cases for AI applications, while 5% say its workforce can use the software for limited purposes. Two percent (2%) report that their institution bans the use of the technology by teachers and staff altogether.
Of those schools that have adopted the tech, more than one in three (37%) reported that they didn’t know how it was being used on campus. Less than a third could articulate how it was being implemented at their schools:
Teaching respondents were using AI to plan lessons (34%), create assignments (30%), create course content (29%), and design coursework (18%).
One in five administrators (20%) use AI to automate or streamline routine activities.
Perhaps most notably, almost one in three respondents (32%) reported that their school wasn’t offering any AI training or, if it was, they didn’t know about it. Further, while 46% of responding schools were offering AI training to their students, only 37% were providing training for their educators, and 27% were providing training for their staff.
The survey results indicate that learners, teachers, and administrators are using AI more frequently in America’s school systems, even though they’ve not been trained on how to implement or manage it. Without that knowledge or skill set, users – and their schools – risk making errors they do not understand that can cause consequences they can’t anticipate.
New Practices Require New Skills
At the same time, across the country, few educational resources are available that provide credentialed AI design and development training. Rather than develop such a program designed to teach AI theory, design, and practice, most schools are simply adopting the AI programming already available through today’s immense digital platforms. And while most standardized AI systems function well for their purpose, they are not designed to address the needs and demands of specific user groups or populations. Instead, what is needed are the introductory and foundational AI courses that generate the AI-based knowledge, comprehension and skill sets that are now so highly in demand.
Two large digital platforms have been actively engaged with schools to develop specialized, AI-focused courses and programs designed to meet both academic and industry standards.
The first ‘official’ AI degree, an associate degree in machine learning and AI, was launched in 2020 through a partnership between Arizona’s Maricopa Community College and the Intel Corporation. Intel adapted its European ‘AI for Youth‘ syllabus as per the specific needs of those community colleges, assisted in training staff and faculty, and was instrumental in its adoption by the schools.
Amazon’s AWS is also actively engaging with academia to both advance AI resources in higher education and overcome inequities found in those systems. Its Machine Learning University has been launched in several minority-serving institutions (MSIs), community colleges, and historically Black colleges and universities (HBCUs) to teach database, machine learning (ML), and AI concepts. The first step in its training program is to educate the educators on the science behind and practices involved in artificial intelligence programming.
At present, California’s UC and CSU systems offer only a few master’s degrees related to AI development; there are no associate or bachelor’s degree paths at present within the state.
AI has erupted into the general public’s consciousness without the slow and measured ‘introduction-education-implementation’ process that typically precedes a whole-community transition from one system to the next. However, the speed at which the tech has been adopted has also eroded the opportunity to ensure appropriate and ethical use of the new digital assets. Further, the unique skills and digital insights needed to design an AI infrastructure are complex and must be mastered if the newly generated system is to replace legacy technology while maintaining or enhancing productivity.