Psychology of Singularity – Your Source for AI News and Information

• Image by Hotpot.ai

Is AI Making Students Lazy and Dumb?

By Jill Maschio, PhD

Teachers across America are concerned about their students’ lack of motivation to work hard when using AI. Students’ test scores seem to be falling, and there is more evidence that intelligence is decreasing. Is AI in the classroom making students less motivated to read, write, and learn or making them more creative and curious about learning?

As AI becomes more popular, we must realize that not all trends are equal. Some trends are helpful, while some are harmful or pretty amusing. We don’t understand the whole story about AI and how it works, according to Sam Altman (Curry, 2024), who addressed questions about AI at an A.I. conference. Whether we know the whole story about AI, education is scrambling to find practical uses in the classroom without first knowing of any real benefit for learning. If you have read my article about AI and its impact on Intelligence, you can access it here.

We need a greater understanding of the potentially harmful and beneficial effects of AI in the field of education. We would assume that there are benefits, much like society and educators thought there would be huge gains in learning when public schools first brought in computers to kindergartner rooms. Later, we learned of some harmful consequences of excessive Internet use. An example of the harm AI might have on academic performance has yet to be fully researched. Marcello Mariana (Dwivedi et al., 2023) writes that generative AI has not only been inaccurate at times but using it leaves out independent thought for humans. Its use can lead to unmeaningful learning for humans and innovation. 

In a study published by the International Journal of Education Technology in Higher Education (2024), researchers studied 494 university students on the effects of ChatGPT and procrastination, memory loss, and academic performance because very few studies have examined such a relationship on academic outcomes. The participants reported on a self-reporting instrument procrastination level and memory loss, such as that they cannot retain information as much as they used to. The researchers collected their latest reported CGA to examine memory loss. They reported memory loss is associated with excessive use of ChatGPT for academic tasks. The continuous use of AI may lead students to develop a lazy attitude toward academic outcomes.

In another study by Petric (2024), students’ grades seemed to improve by using ChatGPT. A closer look at the study reveals that students used the chatbot to prepare for exams and pre-exams, produce entire homework assignments, seminar papers, write part of a project, obtain advice on how to improve a test or assignment, verify the accuracy of one’s work and calculations, obtain critical evaluations of completed assignments, acquire assistance in improving language style and proofreading, etc. These are true indications that students learned course content. This sounds more like when a child comes home and asks his or her parents to help with a science project. The field needs peer reviewed research to show whether using AI fosters learning and intellectual growth. The article in mention here was not published in a peer-reviewed journal. 

Theory of AI and Reduction in Metacognition

I hypothesize that students who lack strong metacognition may not realize what they know and what they don’t while prompting generative AI (e.g., ChatGPT). While people eagerly want information at their fingertips, they can easily ignore the thought process of reflecting on what was learned and what wasn’t to identify what the focus should be on to learn information of content. If this hypothesis is true, then overreliance on generative AI tools for learning may reduce memory formation because active cognitive engagement with information is necessary for selective priority information to be consolidated into long-term memory (Cowan et al., 2021).

The Use of AI may Limit Neurological Activity Needed for Memory Formation

Nature has predetermined that if a specific region of the brain cannot store information, then the brain cannot be modified by experience, leaving the individual unable to remember (LeDoux, 2002). Memory formation is complex. It involves neurochemicals, for instance. When first exposed to stimuli, two neurons will start to fire together, and the neurochemical glutamate helps strengthen the neural connections. If the firing neurons are activated again, the dendrites of the neurons begin to physically grow to establish a neural network.

However, factors may influence the neural network’s stability. According to Shors (2014), processing information deeply can help stabilize neural connections. The duuration of how long a neural network is active may also help stabilize the neurons. For example, the longer we spend engaging with content, the more the neural network stays active – potentially increasing its stability.

So, what is knowledge acquisition?

It is the process of acquiring knew knowledge.  

How is knowledge gained?

Knowledge is gained through the processing of information by encoding it (through sensory systems whether it is something we read or hear). Information must be encoded, organized, and form meaning before going through consolidation to form long-term memory. The working memory is significant to learning to help comprehend new information.  

An excerpt from Kasneci and colleagues’s (2023) article state how AI in the form of LLMs can be used in higher education. 

“ For university students, large language models can assist in the research and writing tasks, as well as in the development of critical thinking and problem-solving skills. These models can be used to generate summaries and outlines of texts, which can help students to quickly understand the main points of a text and to organize their thoughts for writing. Additionally, large language models can also assist in the development of research skills by providing students with information and resources on a particular topic and hinting at unexplored aspects and current research topics, which can help them to better understand and analyze the material” (para. 10). 

They are correct but consider the aspect that by having a LLM write summaries for them, the learner is not engaging in the process of having to figure out what is being said by the author of the text. It is, at times, through the process of trying to determine what is being said and using the working memory to comprehend and form relationships between ideas that greater learning is the end result. And, having to wade through the research to identify gaps and research ideas is part of the learning process. It is what helps turn a novice into an expert. The learner processes information deeply and makes meaning of it, which is one way the brain learns naturally. Having to process information foster independent thought. 

McGehee (2024) examined the correlation between grades and using AI tools in a population of 2,154, students in grades 9 – 12. Students’ use of AI was grouped into using AI as a tool and a facilitator. Using AI as a tool included conducting research and finding information, summarize information, and a writing and editing assistant. AI as a facilitator included explaining concepts and principles into easy to understand language, create study guides and sample test questions, and a tutor or teacher. The research shows a weak correlation with using AI as a tool and a facilitor, but not when using AI for one or the other. The results were for English, health and PE, and social studies, but not significant for mathematics, science, world langues, and other studies. 

The weakness of this study is that it fails to show if learning actually occurred (No cause and effect) or grades were correlated with use of AI when students used it as a means of personal assistance and for writing and editing assistance. Furthermore, the use of AI was not compared to other forms of assistance or methods of studying. Because there was no comparison to other forms of studying or assistance, there is nothing to compare the results to. 

Some Best Practices for Using AI in the Classroom

• The LLM does not triumph over humans in the learning process simply because it can summarize faster.  
• The benefits of using AI in the classroom should outweigh the risks.
• The learner should engage with AI so that deep processing occurs.
• AI should not be used to simply reduce the workload or to ease it, unless the workload is outside of the students’ skill level. For example, if you are teaching a social science course, then having students create a app using an AI app maker where coding is not required, students can have the opportunity to create an app by apply the concepts you are teaching.
• Do not apply technology simply because it is trendy.
• The technology (AI) should give the learner opportunities to think deeply by analyzing, evaluating, and synthesizing information, or creating new information or ideas. 
• The technology (AI) should help foster learning.  

A Theoretical Base for Using AI in the Classroom

There are a number of learning theories. Jean Piaget believed that learning is an accumulation of knowledge where the learner creates their own reality. Without embedding AI into a learning theory, then it’s pretty much hit-and-miss. The educator is simply using another teaching tool without a particular purpose. I propose that educators must use AI (LLMs) with the goal of having students move beyond being a novice and up the continuum towards being an expert. You can read my paper Reassessing Educational Goals for Learning, Assessments, and Outcomes-How Education can Embrace AI for Student Success and Humanity’s Intelligence for more information about that. Humans will need to contain AI in the near future, and part of that entails that AI needs human input.  

When AI Use Influences Loss of Memory Formation

Depending on how a person engages with AI, it may be comparable to using a search engine to locate information that we glance at or to scrolling on our cell phones, both of which may reduce the time spent processing information because if two neurons began firing, the connection would not become stable by moving on to other tasks or scrolling.

There is much more research to consider about this topic. All the research in the literature was not mentioned here, and I’m confident that there is great usefulness for generative AI tools for learning. I do believe that how a person uses AI in the classroom makes the difference, much like my example above about using search engines to locate information. We can have a search engine find information, and then we have a choice of what to do with the information. People can either simply forget it or use the information by processing it further into something meaningful.

Note: A human wrote this essay. 

References

Curry, R. (2024, May 30^th). Sam Altman says OpenAI doesn’t fully understand how GPT works despite rapid progress. Observer. https://observer.com/2024/05/sam-altman-openai-gpt-ai-for-good-conference/

Dwivedi, Y. K., et al. (2023). Opinion paper: “So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy. International Journal of Information Management, 71, 102642, https://doi.org/10.1016/j.ijinfomgt.2023.102642

McGehee, N. (2024). AI in Education: Student Usage in Online Learning
https://michiganvirtual.org/research/publications/ai-in-education-student-usage-in-online-learning/

Kasneci, E., Sessler, K., & Kuchemann, S., et al. (2023, April). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and individual differences 103, 102274. doi: https://doi.org/10.1016/j.lindif.2023.102274 

LeDoux, J. (2002). The synaptic self: How our brains become who we are. New York: Viking Penguin.

Petric, G. (2024). Everyone talks everything with ChatGPT: Students’ uses of ChatGPT an dtheir impact on learning performance. International Journal of Technology and Human Interaction, 20(1). doi: 10.4018/IJTHI.349225

Shors, T. (2014). The adult brain makes new neurons, and effortful learning keeps them alive. Current Directions in Psychological Science : A Journal of the American Psychological Society, 23(5), 311 –318. https://doi.org/10.1177/0963721414540167