We keep on hearing about this ChatGPT. Some are hoping that it is a fad that will go away and others claim that it may take their jobs away. It's not going away and I don’t think your job is going away either.
What is the technology behind ChatGPT?
It is basically a powered large generative large language model (GLLM). Under the hood is an updated version of OpenAI's GPT-3. It has taken the world by storm with its ability to generate human-like text. ChatGPT broke records by reaching a million users in under a week. These models are trained on massive amounts of data and can perform various language tasks such as translation, summarization, and even question-answering. With such impressive capabilities, it's no surprise that people are excited about the impact these models could have on various fields, including education.
One of the biggest limitations of GLLMs is that they are not perfect and can still make mistakes. They are trained on massive amounts of data, but the data is not always accurate or unbiased, leading to the model inheriting these biases. For example, the model may generate sexist or racist text, which could have serious consequences if it's used in education. The tools generate confident sounding outputs, you discern if it is true or accurate. It's essential to be aware of these biases and to take steps to address them.
Another limitation is that these models have a limited understanding of the world. They are trained on a vast amount of data, but the information they are exposed to during training is still limited, in ChatGPT’s case it is a public dataset that reaches 2021. This means that they may not be able to answer more current questions or do tasks based on your private internal systems. Strides are being made to integrate this technology with the current web, internal documents, and other APIs to enhance the experience. For example, a language model trained on scientific text may not be able to answer questions about the arts or humanities.
Given these limitations, it's important to not simply rely on these models to solve all of our problems in education. Yes, I know it has been passing exams, but that just means it is time to adapt. Instead, we need to re-evaluate our approach to education and consider how these models can be used to complement and enhance traditional pedagogical methods. These models can be used as a tool to aid in the teaching process, but they should not replace teachers.
One potential use case for GLLMs in education is personalized learning assistance. These models can generate very good writing that adheres to all the grammatical rules and structures of writing. Each student has their own level of writing skill so it would be nice to have an assistant tailored to the individual needs and abilities of each student. For example, if a student is struggling to come up with a decent intro for an essay, the model can generate a decent starting point that can be edited further. This can lead to a more personalized and effective educational experience for students.
Another potential use case is the creation of educational resources. GLLMs can assist teachers in the creation of lesson plans, study materials, and assessments. This can save teachers a lot of time and effort, allowing them to focus on more critical aspects of education, such as interaction and feedback with students. Furthermore, the model can generate resources that are tailored to the needs of specific groups of students, such as English language learners or students with disabilities.
Language learning is another area where GLLMs can have a significant impact. These models can generate highly realistic examples of how a language is used in different contexts, helping students to understand and master a new language more quickly. Educators again can also generate exercises and assessments to help students practice their language skills. This can lead to a more engaging and effective language learning experience for students.
Don’t Ban It!
History is full of those people that just love to be left behind. You do not want to be that laggard that is stuck in its own ways because of fear. Artificial intelligence is a bit of a mislabel in my opinion. These advancements can only be assistants that are prompted by another human. They do not have a mind of their own.
To wrap it up, the impact of generative large language models on education is exciting, but it's crucial to keep in mind their limitations and not panic!