ChatGPT Changing Education

Background:

ChatGPT Logo

ChatGPT (Chat Generative Pre-trained Transformer) is a state-of-the-art large language model, developed by OpenAI and was released into the public domain on November 30th, 2022. At its core, ChatGPT is a machine learning algorithm that has been trained on a massive dataset of text. This training allows the model to understand and generate human-like text, making it a powerful tool for language-based tasks. One of the key features of ChatGPT is its ability to generate text that is highly coherent and readable. This allows it to understand and mimic the patterns and structures of human language in addition to its ability to answer questions and solve problems. An innovation like this has been regarded as life-changing for many as it is one of the first to democratize access to digital technology and AI. One of the most significant ways that ChatGPT and OpenAI have the potential to disrupt the educational system in both a positive and negative manner. This cutting-edge technology is being used to create personalized learning experiences for students, augment teacher productivity, and even create new teaching methods. While many rave about the potential upside of this technology for education, there are quite a few who are concerned with the repercussions it can have particularly related to its ethical use in education for writing assignments and answering questions juxtaposed with privacy and data protection.

Capabilities and Upside in Education:

ChatGPT is a large language model that has been trained on a diverse set of text data, allowing it to have a broad range of abilities. It can understand and generate text in many languages including English, Spanish, French, German, Italian, Chinese, Japanese, Korean, and many more. It can also perform mathematical operations and answer problem sets related to math. Its features expand over a wide array of math and include descriptions of how to solve the problem by providing a step-by-step walkthrough of each portion of the problem. ChatGPT also confers the ability to answer fact-centered questions in science and other general knowledge topics, but it's not an expert in these areas and will occasionally tell a user when it is unsure of one of the outputs it generated. The model's vocabulary is quite large, with the ability to understand and generate a wide range of words and phrases. It regularly uses 50,000 words in its vocabulary but has the ability to understand inputs to a higher degree. However, this model's knowledge cut-off is in 2021, so it might not have the most recent knowledge on contemporary matters. ChatGPT is a highly versatile chatbot that goes beyond simply mimicking human conversation. Not only can it write and debug computer programs, but it also possesses the capability of composing various forms of written content such as music, teleplays, fairy tales, and student essays. Technology like this has never been seen before in education and has many at the edge of their seats. The ability of ChatGPT to generate human-like text can be used to create personalized learning experiences for students. For example, one professor stated that by having their students write basic functions through ChatGPT they were able to be more productive during classtime and use that time to focus on the discussion portions of the class instead of working on monotonous tasks such as writing essays which could be done and edited at a faster speed.

ChatGPT Solved Student Math Problem

According to NPR, knowing how to use ChatGPT could improve the efficiency and effectiveness of education, as well as make it more accessible to people all over the world. Additionally, OpenAI's research on AI can be used to create new educational tools and resources, such as intelligent tutoring systems which democratizes access to resources that typically cost money and puts them at the fingertips of anyone with an internet connection and computer. ChatGPT and other AI technologies also have the potential to greatly benefit humanity and alter many aspects of our lives, including education, healthcare, and transportation. For example, ChatGPT can be used to create personalized learning experiences for students, which can help to streamline the efficiency and effectiveness of education and make it more accessible to people all over the world, particularly so in communities that have historically lacked resources for quality accurate education. Additionally, OpenAI's research on AI can be used to create new educational tools and resources, and therefore lower the cost and democratize access to heightened opportunities for the future.

Negative Implications for Education:

Ethical Concerns

It is important to ensure that the development and use of ChatGPT and other advanced AI technologies in Education are guided by a strong ethical framework that takes into account the potential benefits and concerns and respects privacy, civil rights, and human dignity as well as plagiarism and inaccurate information. More specifically, measures should be taken into account to prevent the malicious use of these technologies and mitigate the impact of these technologies on plagiarism as this is regarded as an issue in the foreground of education as it pertains to ChatGPT. While ChatGPT has garnered a lot of positive recognition because of the benefits that it can provide to education such as democratizing access to information at everyone's fingertips, there have been significant negative reactions that must be considered. One major concern is the potential for these technologies to be used for malicious purposes, such as spreading disinformation or automating cyber attacks. ChatGPT can be used to generate highly convincing fake news, deep fake videos, and impersonate real people online, which can have serious consequences. One example of this is when ChatGPT was hacked into by a group of self-identified alternative right-wing neo-nazis and used to generate inaccurate responses pushing ideals not based on fact. The difficulty here is that when there is a heavy trust placed on a system like ChatGPT by the moldable minds of young in the K-12 education system, they have no way of qualifying if the information is correct and therefore they inherently assume it is true which has dangerous implications for inciting potential violence within society under the guise of pushing ideals from information that is not accurate. This has the potential to bring significant consequences as an algorithm that hinges on inputs and data that it uses to scrape from different websites will learn information that may be entirely inaccurate in the scope of what it details.

The recent incident of ChatGPT being used to create malware and viruses has raised concerns about the ethical implications of language models in the future and their ability to be trained to do harmful work. This event highlights the importance of ensuring the security of such models and the necessity to be aware of any potential harm when a model may be hijacked to perform the work of a malicious group. It raises the notion requiring one to consider the consequences of releasing a free tool that can perform harmful work and creates a somewhat speculative quagmire of who and what could be liable in the event that someone or something was to get hurt or damaged. As AI technology continues to advance, it will be critical for society to address the ethical concern of the machinery being used to potentially aid in harmful practices such as creating spyware or viruses to gather credit cards or personal information off of the platform.

College and Graduate Schools

Furthermore, while ChatGPT has the ability to be an additive benefit on an assignment, it has been regarded by some universities and professors to retract from student learning because it is able to answer test questions and even perform at a level higher than the average human test-taker such as passing Medical Licensing Exams and the Bar Exam.. The University of California System released a statement and information for how departments and instructors should navigate what they refer to as "the new world" of having technologies like that of ChatGPT in the educational space. They recommend that for the short term, evaluating how pervasive ChatGPT may be for achieving learning outcomes or objectives within the specific courses. This could mean that the educational system is primed for a complete overhaul where most assignments are performed in class or the structure of assignments shifts entirely from any memorization or recitation-centered assignment towards creativity and situationally based where no single assignment is unique or replicable among students and instead is dependent on a created scenario with a unique set of conditions or rules where an AI-based bot may not be used entirely for the assignment. The UC System recognizes that it will be impractical to try to ban or prevent the use of ChatGPT as well as other AI-based tools as "the cat is already out of the bag". In the short term, instructors will need to either make assessments “ChatGPT-proof” or find ways to incorporate ChatGPT in helping students to build new skills. They suggest that in the long term universities as well as their students may need to ponder one of the more difficult questions "if a chatbot can do most of what a college graduate can do, then what is the value of a degree?" . The discussion also focused on how it will be impractical to try to ban or prevent the use of ChatGPT. AI tools are here to stay. "They will improve and become increasingly important across disciplines." In the long run, departments may therefore need to re-evaluate their teaching mission, and ask themselves, if a chatbot can do most of what a college graduate can do, then what is the value of a degree? GSI's at the University of California system said ChatGPT brings promise in the area of automated grading and feedback. ChatGPT could be used to grade assignments, provide feedback on written work, and answer student questions in real time. This can help teachers save time and focus on higher-level tasks, while still providing students with the support they need and focusing more on the connection of faculty to students than on monotonous tasks of grading multiple-choice quizzes. Much of the current discussion in the area of the long-term impact of education is speculative at this point due to how proximal the timeline of the release has been. The impact of what this will have on education is going to continue to change every day.

Inaccuracy

ChatGPT is limited to what it knows and how much it can answer. Since ChatGPT is not connected to a live concurrent database of information, all of its responses are dated to information that was publically accessible and what it was trained on in the year 2021. One of the looming concerns of ChatGPT in education is how the model currently handles commands to which the output is incorrect. The developer, OpenAI, mentions that "ChatGPT delivers nonsensical answers but does so in a seemingly correct tone" and fixing this issue is difficult because there’s no source of truth during reinforcement learning, training the model to be more cautious causes it to decline questions that it can answer correctly, and supervised training misleads the model because the ideal answer depends on what the model knows, rather than what the human demonstrator knows. For example, Stack Overflow, which is a website where programmers can ask questions and seek information needed to restrict users from being able to contribute information that they gathered from the ChatGPT bot. This was due to the issue of there being incorrect answers that were seemingly correct at first. This created a bunch of confusion for rookie programmers who treated much of what they read on this trusted board as fact which led to the dispersal of inaccurate information and many users became frustrated when their designs failed shortly thereafter. While the company did receive a small amount of public backlash for trying to restrict ChatGPT responses on the platform, ultimately many of its users publically acknowledged that it was the correct decision and believe that the risks of potentially incorrect information are too risky for the field of programming at ChatGPT's current abilities.

NPR built on this and has discussed how one potential concern of ChatGPT is the ability to be confidently incorrect on how often individuals should externally fact-check the information it provides and could potentially exacerbate historically inaccurate information because it is only trained through websites and information up to 2021. Critics regard this as potentially negative because for information in scientific and medical fields of education, where information rapidly evolves and there is not an inherently correct or incorrect decision for many scenarios as nothing in the medical field is black and white despite ChatGPT sometimes acting and responding in a binary manner. The high dependency on this type of technology for subjects in the medical education space such as COVID-19 or other medical ailments could create an extraneous amount of liability if the information is inaccurate yet medical students, and future doctors, use ChatGPT in their learning and memorization of scientific facts and learning .

Additionally, there have been academic papers that have developed and correlated certain stylistic patterns in how ChatGPT works in an educational and professional setting. One instance of this was in a legal case ChatGPT may use a framework or a type of provision that is not real or is perhaps less ideal than what a human attorney would think of and develop. This could be problematic because lawyers in training may use this to train under a pretense that is incorrect due to its functional capacity being limited .

In The News

An article from The Atlantic titled "ChatGPT: The AI Writing College Student Essays" discusses the use of the ChatGPT AI language model in creating college student essays. The article highlights how students are using the model to generate personal essays, leading to concerns about academic integrity and the potential for cheating. The article focuses on how the next generations of college applicants are going to have a heavier emphasis on virtual interviews as opposed to simply written text. This was seen as the next step in a way to potentially counterpunch ChatGPT's ability to write an emotionally intense admissions story that is not reflective of the student's writing skills and therefore does not provide an indication of the type of student or person to which the writing represents.

The news story also mentions that some educators are worried that the use of AI-generated essays will lead to a decline in critical thinking skills among students leaving many unable to produce cohesive thoughts under time-constrained pressure without the use of technology if they foster too much dependency on ChatGPT. The author also points out that the use of ChatGPT and other AI technologies in education raises ethical concerns, such as issues of data privacy, and the potential for AI to replace human teachers. Students across the United States have opined that ChatGPT in education is "useful" because it is able to do the ‘soul-searching’ and reflective journaling in a much more convincing manner than them and many of their classmates .

References: