Large Language Model AI Chatbots: Empowering Lifelong Learning Through Digital Innovation

 

By Larry G. Martin, PhD

 

The continuing evolution of Large Language Models (LLMs) provides a generational opportunity for adult learners to incorporate powerful artificially intelligent (AI) chatbots into their digital learning hubs. Some of the most innovative technology companies in the world (e.g., Open AI, Microsoft, Google, etc.) have invested billions of dollars in developing LLMs and generative AI chatbots that can be accessed via computers and cell phones. Since the highly successful launching of ChatGPT-3.5 in November 2022, the availability of AI chatbots has expanded to four more advanced systems (i.e., ChatGPT-4, Claude 2, Bing ai, and Bard ai) that can be considered for adoption by adult learners.

 

Incorporating AI chatbots into a personalized digital learning tool kit is like adding an intelligent digital compass with the ability to understand your individualized personal learning needs and assist you in charting learning pathways to navigate life’s changing circumstances. Notwithstanding the widespread availability of AI chatbots, only about 19 percent of adults have used ChatGPT, and less than 9% have used Bard AI (Business Insider, 2023). Consequently, most adults are not reaping the benefits of this innovative technology. With the ability to interact with people in text and spoken language, these AI chatbots can serve as conversational partners invested in lifelong learning by streamlining the provision of complex and nuanced information.

 

AI chatbots can assist learners to rapidly navigate through vast amounts of data to inform their search for specific knowledge and information. This is a remarkable change from the era of Web 2.0 when adult learners could feel lost as they navigated the vast ocean of knowledge and information available on the Internet. AI chatbots can now provide personalized learning content, digestible explanations, resources, and insights on-demand. However, these systems are in a constant evolutionary churn, which makes it difficult to determine the extent to which they support lifelong learning, the best practices for using AI chatbots, and their limitations and safety concerns.

 

Attributes of LLM AI Chatbots Supporting Lifelong Learning

Like the inventions of writing, the typewriter, and the personal computer, generative AI chatbots (e.g., ChatGPT-4, Claude 2, Bing ai, and Bard ai) were invented for more general purposes; however, they offer powerful attributes as potential self-directed learning partners. These four models were launched in 2023 and share some common characteristics and abilities. All of them were trained on large volumes of digital data, such as books, articles, and web pages, from diverse text and code domains that allow them to make accurate predictions. For example, given the breadth and depth of its training, the excellence of ChatGPT-4’s performance on a wide range of tasks approaches that of an artificial general intelligence (AGI) system (Bubeck et al., 2023). Similarly, all the models have problem-solving abilities to analyze, comment on, and create text and content from diverse data sources; generate and edit text, and engage with self-directed learners on creative and technical tasks. Some can combine different data types, describe images, summarize screenshots, and generate creative content. These AI chatbots can also assist adult learners before, during, and after enrolling in educational and training programs.

 

           AI Chatbots Before Enrolling in Education and Training Programs. Chatbots can serve as virtual assistants for adults considering degree or certification programs. Through application assistance, they can offer guidance on writing effective personal statements, gathering required documentation, and completing applications. ChatGPT-4 can help applicants prepare for entrance exams (such as the GRE, GMAT, or the LSAT), and to practice writing essays. Because they are connected to the Internet, both Bing ai and Bard ai can assist adults with academic and training program research by gathering real-time, up-to-date information on specific educational and training organizations, admission requirements, core courses, electives, and graduation outcomes.

 

           AI Chatbots During Education and Training Programs. As digital assistants, AI chatbots can offer a wide range of academic support to students. All four models can support coursework study through personalized explanations, coursework-related resources and answering study-related questions. Nevertheless, ChatGPT-4 is more likely to create helpful quizzes and flashcards. All the models have the potential to generate ideas and inspire learning by providing fresh ideas when learners are stuck while writing essays. They can also help learners develop their writing skills with personalized feedback on grammar and spelling, structuring essays, citing sources, and evaluating arguments. Claude-2 ai is particularly helpful for creative and literary assignments (Models, 2023). For homework assistance, these chatbots can quickly offer solutions to complex assignments (such as math problems or coding tasks) by explaining the steps involved. ChatGPT-4 or Bard ai can serve as virtual foreign language partners to provide simulated conversations for students to practice their language skills.

 

Using AI Chatbots After Education and Training Programs. After completing degree and certification programs, adults can use AI chatbots as career advancement coaches, virtual partners, and personalized tutors. For career development, Bing ai and Bard ai can serve as career advancement coaches to identify potential job openings, tailor both resumes and cover letters, and assist job interview preparation. For continuing education, these chatbots can provide insights into (and the location of) ongoing learning opportunities such as certifications, seminars, or advanced degree programs. As virtual partners, all models can help identify the required certifications for relevant job markets, generate lists of resources, identify recommended certifications, and assist in updating resumes and LinkedIn profiles. As personalized tutors, these chatbots locate networking opportunities to further career growth by identifying relevant professional associations, conferences, and networking events.

 

Best Practice for Adopting Chatbots

Before adopting LLM AI chatbots, adults should carefully evaluate the extent to which the features and capabilities of the models complement and align with their learning preferences and goals. These AI chatbots are not mutually exclusive. They can be adopted as complementary tools, providing a comprehensive digital assistant ensemble for new knowledge and skills. Because many AI chatbots are financially free, learners can explore their capabilities and suitability by asking sample questions and engaging them in conversations. Learners should consider the model's ability to generate accurate and relevant responses, encourage active engagement, foster critical thinking, and support conversational and personalized learning experiences. By experimenting with different models over time, learners can determine the extent to which they duplicate or complement the features of other models and identify the most appropriate model(s) for their learning goals.

 

LLM Chatbot Limitations and Safety Concerns

As tools in your digital learning toolkit, AI chatbots should be employed as supportive structures rather than standalone teaching devices. They are not substitutes for human intuition and expertise, and they have limitations that should be observed to utilize them safely and effectively.

 

• First, evaluate the datasets upon which the LLMs were trained. Some models were trained on general knowledge, while others used more specialized datasets.
• Second, check if the model is connected to the Internet. Some have real-time Internet connections, while others cannot provide current information.
• Third, the models can produce untrustworthy data. Unreliable outputs are also red flags in academic settings. These organizations require accurate, reliable, and factual content, which some models cannot guarantee (Fernandes, 2023).
• Forth, because they can hallucinate and provide misinformation, AI chatbots should not be used as sole data sources for academic work (Fernandes, 2023). To ensure accuracy, any information generated by AI chatbots should be cross-verified with up-to-date and trusted sources (Models, 2023).
• Fifth, accurate and reliable AI chatbot responses depend on well-designed prompts (i.e., prompt engineering). By providing clear instructions and context, prompt engineering helps ensure the generated content aligns with your purpose (Fernandes, 2023).

 

The widespread availability and generational power of AI chatbots allow all adult learners to break through structural barriers to adult education participation. These chatbots can assist learners in navigating the ever-expanding ocean of digital knowledge by placing in the hands of every adult with a cell phone easy access to the recorded digital history of homo sapiens. The widespread integration of AI chatbots into personal digital learning toolkits and their appropriate use can offer adults flexibility and highlight the learner's control over their learning journey.

 

Up Next: Online Learning Platforms

 

In my next blog post, I analyze online digital learning platforms' key features and capabilities, and which platforms should be considered for adult learners’ digital tool kits.

                                                                                                  

Larry G. Martin, Ph.D.

Professor Emeritus, UWM

Follow me on X (formerly twitter) https://twitter.com/larry_martin29 and LinkedIn https://www.linkedin.com/in/larry-martin-142b528/

 

References

Anthropic. Anthropic: Claude 2, 2023. https://www.anthropic.com/index/claude-2

Bubeck, S., Chandrasekaran, V., Eldan, R., Gehrke, J., Horvitz, E., Kamar, E., ... & Zhang, Y. (2023). Sparks of artificial general intelligence: Early experiments with gpt-4. arXiv preprint arXiv:2303.12712.

Business Insider (2023). There aren't actually THAT many people using ChatGPT. https://www.businessinsider.com/chatgpt-ai-adoption-slow-google-bard-morgan-stanley-2023-6

Fernandes, D. (2023). Why Not to Use ChatGPT for Academic Writing.https://paperpal.com/blog/news-updates/product-updates/why-not-to-use-chatgpt-for-academic-writing

Microsoft. Microsoft: Bing, 2023. https://www.microsoft.com/en-us/bing/do-more-with-ai?form=MA13KP

Models, C. Model card and evaluations for Claude models. https://efficient-manatee.files.svdcdn.com/production/images/Model-Card-Claude-2.pdf?dm=1689034733 Accessed on: (11-18-2023).

Google. Google: Bard, 2023. https://bard.google.com/?utm_source=sem&utm_medium=paid-media&utm_campaign=q4enUS_sem7&gclid=CjwKCAiA9ourBhAVEiwA3L5RFuEHltqo4V_VSoafE84Mz01tBx3bRXQXNFu5tYQTI025o_X2eMbc0RoCBdIQAvD_BwE

 OpenAI. OpenAI: GPT-4, 2023. URL https://openai.com/research/gpt-4

 

Personal Digital Learning Hubs: Transforming Adult Learning and Education

By Larry G. Martin, PhD

The relentless evolution of Web 3.0 innovative digital technologies has created a need for adult learners (regardless of their educational level or employment status) to access and use these technologies for learning and education. Learning is a process that mediates between humans and the societal structures developed by people (Illeris, 2004). Through learning, individuals develop knowledge, abilities, emotions, and sociality, which are essential elements of society's conditions and raw material (Illeris, 2004). Although Web 3.0 is now in its infancy, it is the latest iteration of web-based digital technologies that have disrupted, for three decades, how adults learn, work, and interact. Successive waves of these technologies have shaped our learning, social, and commercial interactions since the emergence of Static Web 1.0 in the 1990s to the current Dynamic Web 2.0 (Saporito, 2023). A major contribution of the Static Web was providing read-only web pages that lacked interactive features but provided quick access to data and information. Although its impact on adult learning and education was marginal, it paved the way for Web 2.0.  

 

The Emergence of Web 2.0

The emergence of the Dynamic Web (2.0) activated core elements of the Static Web that improved the amount and level of user interactivity, social connectivity, and user-generated content that can be accessed globally (Saporito, 2023). It unleashed tremendous changes as workplaces, commercial enterprises, and learners adjusted to a new digital world. Adult learners can now conveniently and rapidly access a wide variety of digital content through mobile internet and social networks, powerful mobile cellphone devices, digital apps (e.g., Facebook, Twitter [now X]), and others (Saporito, 2023). Using 5-G technology, adult learners can communicate with each other and instructors, create content, and access various digital learning tools (e.g., web pages, social media networks, learning management systems, online learning platforms and learning apps, books, articles, news, etc.). However, little was known about how adults have adopted and used Web 2.0 innovative technology tools for personal and job-related activities. Therefore, a 2015 Pew Research Center investigation surveyed a national sample of 2,752 U.S. adults. It found most adults professed to have lower levels of digital skills and trust in the online environment (Horrigan, J. 2016). 

 

Over half (52%) of U.S. adults were unlikely to use digital tools in their learning (Horrigan, 2016). Most non-users were women, adults aged 50 and older, racial minorities, people living in lower-income households, and individuals with lower levels of formal education. A middle group of about 31 percent of adults reported having technology resources available and acknowledged the ability to use the internet to pursue their learning interests confidently (Horrigan, 2016). However, these adults reported they were not energetically involved in electronic learning at a high level. Only 17% of U.S. adults expressed confidence in their abilities to use digital tools to pursue learning and were actively engaged as learners (Horrigan, 2016).

 

These survey results suggest that the Web 2.0 digital revolution may be viewed as intimidating and a cause for apprehension among adults who currently (or may soon) find themselves possessing irrelevant skills and abilities in a new and evolving digital economy. However, like two sides of a coin, this revolution also creates digital learning tools and opportunistic pathways for adults to continue learning. An increasing list of digital learning tools is now available to provide adults with free and subscription-based access to previously unreachable levels of knowledge and skills. Adults can now harness the power of automation and connectivity to amplify their performances in a wider range of contexts. By mastering digital learning tools, this tumultuous change can be transformed into an empowering journey, leading to a new era of increased opportunities and potential as we witness the rise of Web 3.0.

 

The Emergence of Web 3.0

Web 3.0 continues to emerge during the 2020s, and it is overhauling and modernizing our learning and educational tools as it is simultaneously transforming our workplaces, classrooms, and broader communities. Web 3.0 also connects to and integrates four emerging computing technologies: spatial technologies (such as artificial and virtual reality); physical technologies (e.g., Internet of things); cognitive technologies (e.g., artificial intelligence); and distributed technologies (e.g., Blockchain) (René, 2019). Some of its key features include:

 

1. Decentralization of web-based information and its storage in multiple locations simultaneously. 
2. Trustless and permissionless transactions that allow users to interact directly without an intermediary or permission authorization from a governing body.
3. Natural language data processing of textual data.  
4. Connectivity and ubiquity allow information and content access by multiple applications and an increasing number of web-connected everyday devices (Saporito, 2023).

 

A more fully developed Web 3.0 will employ artificial intelligence and machine learning information processing to produce faster and more relevant user results.

 

Web 3.0 transformations are expected to replace our traditional and comfortable workplace knowledge and educational practices and reforge them entirely. Although decentralizing the web and permissionless systems will give users much greater control over their personal data, it also carries legal and regulatory risks (e.g., increases in cybercrime, hate speech, and misinformation). For example, in 2018, Microsoft employed over 800 editors to help select and curate news stories shown to millions of readers worldwide as they login to devices using Microsoft software (O’Sullivan & Gordon, 2023). In recent years, Microsoft laid off editors and replaced them with Artificial Intelligence (AI) chatbots to scrape the web for newsworthy stories shared with readers. This strategic decision is being connected to the websites’ recent proliferation in the publishing of false and bizarre stories (e.g., false claims that President Joe Biden fell asleep during a moment of silence for victims of the Maui wildfire) and unprincipled actions (e.g., the creation of an auto-generated poll asking readers inappropriate questions after reporting human tragedies), and amplifying stories published on obscure fringe outlets (O’Sullivan & Gordon, 2023). This scenario of replacing some humans with technology while retaining others to manage these technologies is expected to play out in numerous workplaces as Web 3.0 technologies are improved and adopted by employing institutions and organizations. By using digital learning tools located in personal learning hubs (i.e., a digital learning tool kit), adults can upskill and reskill to stay ahead of the arc of Web 3.0 change.

 

Web 3.0 Digital Learning Tools and Resources

Personal digital learning hubs are spaces created on cell phones, computers, and digital tablets that provide 5-G internet-connected technology-rich learning environments with both physical components and virtual apps that provide formal and informal learning opportunities. The thoughtful creation of hubs comprised of targeted Web 3.0 digital technologies to address their learning needs has the potential to provide accessible, affordable, and private means for adults to stay more easily engaged as lifelong learners, workers, and active citizens. However, a dizzying array of digital tools are available, and each has distinctive characteristics and attributes that might be helpful to adult learners.

 

In this online forum, I will identify and describe the primary attributes of these technologies that make them attractive for adult learners, how they are used for education and learning, how they currently influence adult learning, and the key limitations that might inhibit their use. Examples of these technologies include (but are not limited to): large language model AI tools; online learning platforms; mobile elearning platforms and gamification; immersive augmented and virtual reality; the Internet of things and wearables; learning analytics and big data; smart adaptive learning management systems; blockchain technology; and others (see figure 1).

 

Figure 1: Personal Learning Hub of Web 3.0 Digital Learning Tools

Building a highly integrated and interconnected personal technology hub is essential for lifelong learners seeking to identify, store, and deploy digital teaching and learning tools and resources. An in-depth analysis of the different categories of Web 3.0 teaching and learning technologies would allow learners to identify which digital tools and systems can be helpful and adopted for their toolkit to achieve specific learning goals.

 

Up Next: Large Language Model AI Tools for Lifelong Learning

 

In my next blog post, I analyze the key features and capabilities of four large language models (i.e., ChatGPT, Claud, Bing, and Bard) that can be used as AI digital learning assistants.

 

Larry G. Martin, Ph.D., Professor Emeritus, UWM
Follow me on X (formerly Twitter) https://twitter.com/larry_martin29
LinkedIn https://www.linkedin.com/in/larry-martin-142b528/

References

 

Horrigan, J.B. (2016). Digital Readiness Gaps. https://www.pewresearch.org/internet/2016/09/20/digital-readiness-gaps/

Illeris, K. (2004). The three dimensions of learning: Contemporary learning theory in the tension field between the cognitive, the emotional and the social. Roskilde University Press.

O’Sullivan, D. & Gordon, A. (2023). How Microsoft is making a mess of the news after replacing staff with AI. https://www.cnn.com/2023/11/02/tech/microsoft-ai-news/index.html

René, G. & Mapes D. (2019). The Spatial Web: How web 3.0 will connect humans, machines and AI to transform the world‖ Paperback.

Saporito, S. (2023). Web 3.0 Explained, Plus the History of Web 1.0 and 2.0. https://www.investopedia.com/web-20-web-30-5208698

 

Welcome to the Online Forum!

Dr. Larry G. Martin

This online forum is moderated by Dr. Larry G. Martin. Dr. Martin is Professor Emeritus of Adult and Continuing Education Leadership at the University of Wisconsin-Milwaukee (UWM). As a faculty member for over 36 years, he routinely taught graduate courses on program planning in adult education, administration of adult education programs, evaluation of adult education programs, and others. 

A 2015 inductee into the International Adult and Continuing Education Hall of Fame, he has published and co-published eight edited books, numerous articles, and book chapters. For example, he co-edited an international compendium (Vols. 1–4) with Alan B. Knox and Simone C. O Conceição, entitled Mapping the Field of Adult and Continuing Education: An International Compendium. The purpose of this Compendium was to produce and encourage the use of a major reference work based on mapping the knowledge base of the adult and continuing education field. More recent publications include Black Men and the Digital Divide (2016) with Simone C. O Conceição and Politics, Policies, and a Human Rights Agenda for Racialized Minorities: The Role of Adult Education (2021) with Danielle Apugo. 

Dr. Martin's interest in the vast potential of Web 3.0 innovations for teaching and learning has led him to actively develop scripts and produce 18 podcast episodes on innovative digital technology, adult learning, and education via the Adult Learning Exchange Virtual Community Podcast. He earned a Ph.D. at the University of Wisconsin-Madison.