Empowering Special Education Faculty: Navigating the AI Landscape in Higher Education for 2023-2024.
Authors: James D. Basham, Ph.D., University of Kansas, Matthew T. Marino, Ph.D., and Eleazar Vasquez III, Ph.D., University of Central Florida; info@ciddl.org
The CIDDL team wanted to provide you with some thoughts on the growth and adoption of Artificial Intelligence (AI) as we begin the 2023-2024 academic year. We have been working to identify areas of consideration as higher education faculty navigate the future of AI and utilize its potential to support individuals with disabilities. The recent public release of large language models, such as ChatGPT and Google Bard, did not happen overnight. AI in various forms has been around for decades. One of CIDDL's first blog series was on AI (see Marino, 2021). From video games to personalized learning, many on the CIDDL leadership team have been integrating AI into their research agendas for years. However, recent advancements in AI dramatically advanced the technology, which has foreseen consequences across society. Because of the increased availability of data, enhanced processing speeds, and improvements in algorithms, AI is growing at an exponential rate (Bongonmin, et al., 2020). The purpose of this article is to help inform conversations and provide considerations for advancing the understanding of AI.
What are some key terms and concepts associated with AI?
Let’s start with a general overview of how Large Language Models (LLMs) process data, such as text and images, using a neural network. The network is organized into multiple nodes and layers, which are used to continually adjust how the analyzed data is presented. What makes this so revolutionary is the fact that these LLMs can learn from, and improve their responses, based on their experiences. If the model runs ten times, the response on the 10th trial will be more accurate and efficient than the first trial. Additionally, they are constantly self-analyzing and self-correcting. Finally, LLMs look to humans to enhance their performance through a process called, “reinforcement learning on human feedback” (RLHF). During this process trained supervisors and end users rate LLMs performance, ranking the responses and noting when they are inaccurate.
There are a few more acronyms to understand. Ever wonder what GPT stands for in ChatGPT? Generative Pre-trained Transformer. “Generative” means ChatGPT can create a new text based on a predictive model. It looks across the large database it was trained on and generates the most logical next word. “Pre-trained” comes from the world of machine learning. It means ChatGPT has been trained about how sentences are formulated and will choose terms conforming to the parameters of the written language. “Transformer” is a reference to the overall model. Think about it as a master plan for an automobile, describing how interconnected parts in the model interact.
Finally, what does API stand for in Open AI API? Application Programming Interface. Application refers to software with a specific function. An API is defined as a set of rules allowing different software applications to interact with each other. It provides protocols for enabling data transfers between systems, which enables multiple stakeholders to use the data across a diverse array of analyses. APIs allow the integration of disparate applications in order to improve efficiency, flexibility, and innovation.
What does exponential growth in AI mean?
Exponential growth is a mathematical concept that describes a rapid and consistent increase in quantity over time, wherein the rate of growth is proportional to the current quantity. In other words, as the quantity increases, the amount by which it grows also increases, leading to a progressively steeper upward trajectory. Exponential growth in AI refers to the rapid and accelerating advancement of AI technologies and capabilities over time, doubling at rapid and consistent intervals. An important property of exponential growth is it can lead to very large values in a relatively short period, especially in comparison to linear growth, where the growth rate is constant but the absolute increase remains the same over time. To see a timeline visit There’s An AI for That (https://theresanaiforthat.com/) to see the growth of AI applications since 2015 and note the rapid increase in development.
What does this mean for those who research and prepare people to serve individuals with disabilities?
This means the power and presence of AI are rapidly changing the world, and the field of special education is no exception. Marino et al. (2023) provided an in-depth overview of the multiple considerations for how AI will change the field. To support this understanding, here’s a snapshot of topics CIDDL leadership is actively engaged in or is monitoring. If they have not already, these topics should start playing a role in your department’s discussion about the future of special education, early intervention, related services, and personnel preparation.
- Equity and Inclusion: AI has the potential to both support and challenge inclusivity in education. Researchers should critically examine how AI tools can address learning disparities and provide tailored support for students with disabilities. They should also be cautious about potential biases and ensure AI technologies do not perpetuate inequalities. A fear is school teams simply believe what they are told without question, which is essentially going to hand decisions over to the machine because AI will see things humans do not. The best approach is an educated human and machine convergence model, wherein the AI is supporting human decision-making. However, this going to require personnel who understand how to appropriately engage and integrate AI across various situations and contexts.
- Personalized Learning: AI can enhance personalized learning experiences for ALL students, especially those with disabilities and other diverse learning needs. Soon applications will be able to adapt content, pacing, and teaching strategies based on individual moment-to-moment engagement. Researchers should explore how AI can optimize individualized education programs (IEPs) and offer real-time interventions based on student progress and needs. For instance, Basham et al. (2016) found when designed appropriately personalized learning can support better-than-expected learning outcomes for students with disabilities. However, the adoption of a personalized education system is completely disruptive to the current bifurcated (special) education system existing today.
- Teacher Professional Development: The integration of AI in K-12 education will require significant teacher professional development. Researchers should investigate the training needs of special education teachers to effectively use AI tools, and assess how AI can complement their expertise rather than replace their roles. For instance, consider how personnel should be prepared to support personalized learning environments (see Basham et al, 2016).
- Ethical Considerations: AI is going to introduce a host of issues associated with equity and inclusion. While much of the current work in the field is focused on the efficacy of interventions, the age of AI is going to introduce a host of additional issues that special education is going to struggle to answer. Special education researchers should address ethical concerns related to AI implementation, such as accessibility, data privacy, informed consent, transparency, and appropriate use. They should advocate for policies that safeguard students' rights and ensure the responsible use of AI technologies in educational settings.
- Collaboration and Interdisciplinary Approaches: The development and implementation of AI in special education requires collaboration between researchers from diverse fields, including education, computer science, psychology, and ethics. Special education researchers should engage in interdisciplinary discussions to contribute to the design, evaluation, and improvement of AI tools for students with disabilities. Consider the role interdisciplinary knowledge, skills, and collaboration play in preparing the future leaders of the field. For instance, Zhang et al., (2020) identified various interdisciplinary research considerations related to personalized learning.
Where can you start?
By staying informed about these critical aspects, professors and researchers in the field can help ensure beneficial outcomes for students with disabilities as it relates to the integration of AI. Here are five things you can do to prepare yourself for the future.
- Learn about AI yourself. This includes understanding the basics of AI, as well as the potential benefits and risks of using AI in education. Of course, you can simply “Google it”, review the previous content on CIDDL, or contact us for support.
- Talk to educators about AI. This can help you to get a sense of what other educators are already doing with AI, as well as their thoughts on the potential of AI in education. Consider partnering with districts to help them better understand AI.
- Explore AI tools and resources. There are a number of AI tools and resources available for educators, so take some time to explore what's out there. CIDDL has already pointed out a number of tools and resources you can try.
- Experiment with AI in your own teaching. This is the best way to learn how AI can be used effectively in the classroom. We encourage you to try new things, especially in your teaching. Talk with your students about what worked, what did not, and why.
- Share your experiences with others. This can help to build a community of educators who are interested in using AI in education, and it can also help to spread knowledge about the potential of AI in education. Share your experiences in the CIDDL Online Community. Additionally, if you to be more engaged apply to join the CIDDL AI Workgroup.
It's important to note that while AI is experiencing exponential growth in certain aspects, it still faces challenges and limitations, such as ethical concerns, biases in data, and the need for more explainable and interpretable models. The growth of AI is a complex interplay of technological, scientific, and societal factors that continue to shape its trajectory. Learn more about AI and engage in critical conversations now, rather than waiting for AI to drastically impact the field. Also know, CIDDL is here to help. If you want to learn more, have suggestions for content, or have questions let us know.
Join the Conversation
With that in mind, there are a lot of really helpful and interesting AI solutions emerging, with more coming out every day. Which have you found the most helpful? Which would you or would you not recommend? Head over to our community and share with us!
References:
Basham, J. D., Hall, T. E., Carter Jr, R. A., & Stahl, W. M. (2016). An operationalized understanding of personalized learning. Journal of Special Education Technology, 31(3), 126-136.
Bongomin, O., Gilibrays Ocen, G., Oyondi Nganyi, E., Musinguzi, A., & Omara, T. (2020). Exponential disruptive technologies and the required skills of industry 4.0. Journal of Engineering, 2020, 1-17.
Marino, M. T., Vasquez, E., Dieker, L., Basham, J., & Blackorby, J. (2023). The future of artificial intelligence in special education technology. Journal of Special Education Technology, 01626434231165977.
Marino, M.T. (2021, March 6) AI Episode 1: Intro To Artificial Intelligence In Teaching. CIDDL. https://ciddl.org/artificial-intelligence-in-teaching/.
Marino, M. T. (2021, April 15) AI episode 2: What does an AI teaching assistant look like? CIDDL. https://ciddl.org/ai-episode-2-what-does-an-ai-teaching-assistant-look-like-2/
Murugesan, S., & Cherukuri, A. K. (2023). The Rise of Generative Artificial Intelligence and Its Impact on Education: The Promises and Perils. Computer, 56(5), 116-121.
Zhang, L., Basham, J. D., & Yang, S. (2020). Understanding the implementation of personalized learning: A research synthesis. Educational Research Review, 31, 100339
Clarke, L. (2023). Call for AI pause highlights potential dangers. Science, 380, 120–121.