Education

Promises

NLP in combination with other extended reality technologies can enhance the learning experience by offering more engaging delivery of lessons.
With the help of NLPs, teachers can efficiently accommodate assignments to fit the different aptitudes of their students.
Unlike a human educator, a chatbot is able to repeat instructions exactly a large number of times and without losing its temper.

Already existing examples of NLPs in education are ChatGPT (Open AI) and other virtual assistants.
NLPs can be used to improve the overall quality of academic writing through grammar and syntax checks.
They can help reduce the workload of both teachers and students.
Conversational agents are often used in the education of autistic children or in the rehabilitation of disabled people as well as teaching foreign languages through Intelligent Languages Tutoring Systems (ILTSs).

Datasets based on human educators also risks importing the undesirable traits into the conduct of the chatbot.
Dialogues with conversational agents are recorded. When a chatbot converses with vulnerable people or children, these records can contain sensitive information. It is important to include the collection, storage and use of these traces in a legal framework.
Children are naturally inclined to talk to inanimate objects such as toys, an even stronger attachment is formed when they can respond and interact. How will the implementation of chatbots in everyday gadgets and toys affect the development of a child?
Intentional or not, plagiarism can surface in a document crafted with aid of NLPs. Moreover, how can schools define and verify integrity and cheating on the students’ assignments?

Chatbots perform better or worse depending on the size of the data set that has been used to train the large language model. But even within a well-represented language such as English, there are differences in performance depending on slang or dialect. NLPs capture more accurately the language use of certain groups, which could lead to discrimination of students not belonging to these groups.
Sentences taught by a chatbot may be too literal and lack nuance.
A conversational agent may teach the student to pronounce sounds inhumanly, based on statistical averages of tone, energy and rhythm.