Photo by Vlado Paunovic + thisisengineering pexels
AI &
Learning Gap
IDT200x
At ATXPO, an academic tech expo for Bay Area colleges I recently attended, instructional designers attempted to come to terms with their own relationship with Artificial Intelligence (AI). Practitioners provided approaches to checking the pulse on student AI integration in their learning process.
This moment leads me to personally confront AI by exploring its current manifestations leading to my own reckoning with the latest technology.
In the Instructional Design Document (IDD), I will address a learning (or knowledge) gap that my proposed minicourse, Bento Journey, hopes to address.
Background photo by Scott Webb
Artificial Intelligence Overview
Artificial Intelligence in the context of learning encompasses a host of technologies that enhance the learning experience. AI applications collect user information (ex. assessments), customize responses (ex. feedback on inputs), and provide comprehensive data analysis on student performance. It further automates processes from grading to content creation to even tutoring. Such technologies include machine learning, natural language processing, and Intelligent Tutoring Systems. The following are a few key areas for growth within the A.I. realm of learning experience design.
Key Areas for Artificial Intelligence
Content Personalization
Learner Centered Design
AI can amplify learner centered design through customization considering:
student need,
skill level,
preferences,
interests,
enabling more personalized instruction. IDs can configure course outcomes to apply to more specific learning requirements (job training, performance expectations, goals, skills, etc.) Collected information about learner abilities and progress enable customization.
Adaptive Learning
is one option for customization. This technique adjusts course materials according to learner performance.
Learning Experience Development
Content Creation
Brainstorm ideas, Generate content
Generative AI takes course designer's text prompts related to desired course outcomes and creates engaging course content with ID as curator/editor
Personalization
Content, as well as quizzes, flashcards, summaries specific and relevant to learner are auto generated
Assessment
Increase Formative
AI generated assessments along the way provide immediate feddback to learners prior to being graded. This lower stakes feedback mechanism enables learners to learn from mistakes without point deduction.
AI can automatically grade assignments (lower stakes), essays and multiple-choice questions. Instructors can provide more coaching during formative stages of assessments. Analytics can inform instructors where
Real Time Support
Intelligent Tutoring Systems
provide personalized feedback based on performance. Can be used toward mastering a skill or concept. Can be like a walking encyclopedia that provides basic academic information that deconstructs concepts with examples on demand. Response can be human-like. Can tailor to individual preferences.
Natural Language Processing
is defined as conversational interfaces, like chatbots and virtual assistants who are there to answer your questions, customer service style. The AI can discern meaning from human language through reading comprehension.
Multiple Modalities
Accessibility
AI enables multiple modalities to be autogenerated. Current assistive technologies that help support visually, aurally, and physically impaired, and second-language users include image-description generation, automatic speech recogntion (ASR), enabling close captioning and transcripts on videos.
The multiple modalities further support neurodiverse learners
Simulation-based Learning
Resource Allocation
Simulation-based learning coincides with other topics above such as Intelligence Tutoring Systems and simulation games. Medical simulations, smart edutainment, and virtual reality and virtual campus are enabling remote, personalized learning. Instructional simulations enable one to envision what work environments and problem solving may look like in professional environments.
Universe by ViewSonic is one such virtual campus for online learning similar to Metaverses Decentraland and SecondLife...
Data Analytics
Predictive & Learning Analytics
Predictive Analytics
Data from user interactions analyzed by AI algorithms identify knowledge gaps. AI's prediction models adapt learning pathways to retain potentially disengaged learners optimizing course design.
Learning Analytics
assist instructors in tracking learner progress and identify learning gaps. Course material adjusts to learner rather than vice versa. Then student performance across the course can be tracked over time for course design improvement.
Gamification
Extrinsic motivation
Taking cues from the popularity and expansive presence of video games, instructional designers can increase learner engagement with familiar cues. Extrinsic motivators include gaining rewards, achievement in difficulty levels, and events prompted by progress points.
Educational games
Puzzles, word-matching, or role playing games provide more engaging, interactive, non-linear way of testing knowledge
Background photo by Miguel á Padriñán
Content Personalization
Adaptive learning advantages
-
Helps learning retention and engagement
-
Improves time efficiency of training
-
Tracks learner performance over time
-
Inputs multiple data points such as, time on task, level of proficiency required, and answer choice, for more "intelligent," dynamic response
Examples
-
If a student answers a question correctly, more difficult or complex questions are posed.
-
Providing data on student progress and performance helps instructors adjust teaching or next online lesson according to individual results.
Personalized learning advantages
-
Supports learner engagement
-
Instant feedback enables learner to improve
-
Enhances experience with positive feedback based on actual performance (Gamification)
Personalized learning examples
-
Instant feedback on performance
-
Provide encouragement incorporating positive feedback where the student performed strongly
Real Time Support
Intelligent Tutoring Systems
App acts as host to learning activities with feedback and self-learning activities at key points. Like ChatGPT but embedded in the learning experience design.
Intelligent Tutoring Systems Examples
-
Initial assessment on subject matter knowledge
-
Knowledge gaps identified lead to learning pathway that bridges current level with desired outcome.
-
Provides relevant exercises with immediate feedback
-
Adapts to improvement with next level content
Natural Language Processing
A more conversational format guides the participant as part of the learning experience. AI directs learner to further information.
Natural Language Processing Examples
-
NLP algorithms can recognize and code different information into categories from a spoken report that it translates into structured data such as electronic meidcal records.
Data Analytics
Predicative Analytics
This integrated approach improved student engagement, collaboration, and satisfaction. Integration means using AI performance predication model combined with embedded learning analytics (Gibson).
Predicative Analytics Examples
-
Identify student behavior patterns which helps build predictive models using machine learning.
Learning Analytics
Instructional designers can use overall trends in learner data from learner analytics to create more effective and up to date learning experiences based on demographic's needs.
Learning Analytics Examples
-
AI records multiple metrics: scores, completion time, attempts made, even learner engagement by number of times logs in and amount of time spent on each activity
-
Insight into learning behaviors can predict future performance and enable learning interventions when necessary
Learning Experience Development
Content creation examples
-
AI generates smaller chunks of information in the form of text summaries
-
Learner data provides inputs for chatbots and virtual assistants to deliver personalized learning content
-
Diagnostic assessment enables AI to adjust learner's starting point on course and broaden learning resources on challenge topics
-
Relevant tutorials, activities including more interactive modalities, and infographics could enhance learner experience and add to comprehension
-
Overall pattern of student body's performance can inform future iterations of the course
-
Most or least used resources could further inform course revision
Most of all, AI-powered content generation saves time by speeding up overall course development. It will be easier, faster, and more flexible in generated solutions. The instructional designer would act as initiator, data analyzer, content curator and editor in this scenario.
Multiple Modalities
Learning Experience = User Experience Design
When instructional design meets user experience design, learning experience design happens. User experience design places the user or learner at the center of the design process. Much of what I learned from Google User Experience Design can be repeated here. When creating learning experience design, the objective is to be as inclusive as possible. The online experience includes a global audience, which may include an audience that has limited technological access, may be acquiring digital literacy. English may be a second or third language.
According to the World Health Organization, 16% of the global population, or 1.3 billion people experience a significant disability today. Furthermore, according to the Zurich Insurance Group, 15-20% of the world's population is thought to be neurodiverse (dyslexia, dyspraxia, ADHD, and autism).
Online learning has further provided access to courses that previously weren't accessible before for geographic, financial, and time-challenged students. It as enabled lifelong learning and career update opportunities as never before. AI has further enabled access on mobile devices, which are more widely distributed worldwide. According to trend forecaster, Exploding Topics, 95.3 % of reporting users access the Internet via mobile phones, with 94% using smartphones vs. 63.4% using laptops or computers.
Gamification
This behaviorist approach incentivizes learner through external rewards. AI enables a personalized gamified learning experience.
Examples of AI-based gamification
-
language-proficiency assessment provides appropriate language level for learning
-
AI tracks learner progress and provides responsive difficulty level, according to whether learner is struggling or able to get through the activity
-
Learner activity can prompt relevant puzzles or word matches to increase enjoyment of learning
-
Instant feedback on progress through badges, points, or leaderboard can further incentivize learning to keep going
Assessment
Examples of Assessments
-
Multiple choice is currently widely used in MOOC assessments
-
AI goes beyond spellcheck and can customize a pathway according to an essay format of assessment: grammar, sentence structure, word choice. NLP enables more complex assessment with checks for coherence, thesis statement, and support with evidence
-
AI provides immediate feedback with writing strategies that could improve the learner's essay
-
AI tracks progress and advances content according to learner performance and readiness according to assessment results
Simulation-based Learning
A representative from Universe came to the ATXPO, a Bay Area academic tech expo, I attended. Universe is a virtual campus where students and teachers create avatars and meet in a Metaverse-like 3D-rendered space.
Teachers can design and create a 3D classroom space. Where the 3D space's strengths come in appears to be student attendance and participation. The 3D classroom has a discussion mode where students can interact and "mingle" with one another. In lecture mode, you can see if students are paying attention, as in a regular classroom. Instructors can further monitor breakout rooms to see whether discussions are actually taking place. In breakout rooms, smaller groups can share presentation builds and respond to one another's inputs with reactions. You can also administer a quiz on the dashboard with instant feedback on who responded correctly and who didn't.
Devlin Peck in his video, Is the Metaverse the Future of Learning? the potential of similar spaces like the Metaverse, Decentraland, of recreating historical environments to travel to help illustrate literature or history. For now in Universe, the classroom appears to be a set of skins to choose from. I believe sites like these are meant to gameify and engage younger audiences that are accustomed to working online. Horizon does something similar for the corporate environment.
Works cited
“12 Generative AI & Ai Resources.” Rasa.Io, 26 July 2023, rasa.io/pushing-send/ai-resources/.
“Artificial Intelligence in Instructional Design.” edX, learning.edx.org/course/course-v1:USMx+LDT200x+3T2023/block-. Accessed 22 Oct. 2023.
“Disability.” World Health Organization, World Health Organization, 7 Mar. 2023, www.who.int/news-room/fact-sheets/detail/disability-and-health.
Gibson, Rob. “10 Ways Artificial Intelligence IS Transforming Instructional Design.” EDUCAUSE Review, 14 Aug. 2023, er.educause.edu/articles/2023/8/10-ways-artificial-intelligence-is-transforming-instructional-design.
Howarth, Josh. “Internet Traffic from Mobile Devices (Oct 2023).” Exploding Topics, Exploding Topics, 28 Sept. 2023, explodingtopics.com/blog/mobile-internet-traffic.
Kereselidze, Manuchar. “The Role of Artificial Intelligence in Instructional Design.” eLearning Industry, 11 Aug. 2023, elearningindustry.com/role-of-artificial-intelligence-in-instructional-design.
“Learn Prompting: Your Guide to Communicating with AI.” Learn Prompting Your Guide to Communicating with AI RSS, learnprompting.org/. Accessed 23 Oct. 2023.
MIT Technology Review, www.technologyreview.com/. Accessed 23 Oct. 2023.
Peck, Devlin. Is the Metaverse the Future of Learning?, 28 Dec. 2021, https://youtu.be/q9HTnWGdes8?si=0M8EkhWDokXPPqs8. Accessed 22 Oct. 2023.
“The NCES Fast Facts Tool Provides Quick Answers to Many Education Questions (National Center for Education Statistics).” National Center for Education Statistics (NCES) Home Page, a Part of the U.S. Department of Education, 2021, nces.ed.gov/fastfacts/display.asp?id=80#.
“Universe by Viewsonic.” UNIVERSE by ViewSonic, universe.viewsonic.io/. Accessed 22 Oct. 2023.
Implications for Instructional (Learning) Design
Will Artificial Intelligence replace us as instructional designers?
According to Gibson, AI won't replace us at present because learning is still a personalized and humanistic, by which I believe he means, interpersonal experience. However, we do need to come to terms with it and select appropriate AI for course creation.
In other words, we need to maintain our sense of authorship by using AI as tools and optimizing their adaptive strength through becoming more adept in applying the tools where needed.
Given the widespread use of blended learning in K-12 classrooms, the next generation is accustomed to some level of hybrid learning. Further, online learning has been a mainstay of higher education. "In fall 2021, some 9.4 million students, or 61 percent of all undergraduate students, were enrolled in at least one distance education course"(National Center for Education Statistics).
Therefore AI has the potential to provide efficient, personalized learning experiences that provide learners with instant feedback and more conversational style engagement and relevant content to the student's level. Kereselidze believes, rather than replacement, IDs roles will evolve into consultants of AI solutions ensuring content aligns with learning objectives. Learning institutions will curate learning content ensuring relevance and accuracy and alignment with school's mission.
How to keep up with AI
Here are some good sources to follow when it comes to AI
I found this resource rich list of AI experts to follow, not surprisingly, MIT is among them. I picked a few from the list I intend to follow:
-
MIT Technology Review – This review so far seems to look at the social issues around AI, amongst other technology related topics.
-
Learn Prompting – If you're in need of learning more about how to communicate with AI, this would be a key resource for getting started. And it's free.