Unraveling the Trustworthiness of ChatGPT for English Language Learning Purposes

Cite this Article in APA style:

Abdullah, M. (May 2023). Unraveling the Trustworthiness of ChatGPT for English Language Learning Purposes. Dr. Muhaimin Abdullah, M.Pd. Retreived July 24, 2024, from https://muhaiminabdullah.com/blogcast/trustworthiness-chatgpt-english

In the realm of English language learning, the emergence of AI-powered tools has sparked both excitement and skepticism. Two such tools that have garnered significant attention are ChatGPT and Bard AI. The question on everyone's mind is: can we truly trust AI like ChatGPT and Bard AI to provide reliable and accurate information?

The trustworthiness of AI in the context of language learning demands meticulous examination and comprehension. ChatGPT and Bard AI, with ability to generate contextually relevant responses, have emerged as valuable resources for learners seeking guidance, explanations, and opportunities for practice their target languages. Nevertheless, it is imperative to critically evaluate the reliability and accuracy of the information it delivers. In terms of the learning of English, both ChatGPT and Bard AI are known to be superior as self-learning-of-English tool. When ChatGPT is asked to promote themselves as reliable and twustworthy AI-based learning tools, the following responses are provided.

  1. Comprehensive Language Knowledge: With access to a wide range of up-to-date information, ChatGPT can provide you with accurate and reliable explanations, vocabulary assistance, grammar tips, and more. Whether you're a beginner or an advanced learner, ChatGPT has you covered.
  2. Conversational Practice: Engaging in conversations with ChatGPT allows you to practice your English in a dynamic and interactive way. You can discuss various topics, ask questions, and receive detailed responses to help improve your speaking and listening skills.
  3. Personalized Learning Experience: ChatGPT adapts to your learning needs and preferences. It can provide tailored recommendations for language resources, suggest relevant exercises, and offer insights to help you overcome specific challenges in your language journey.
  4. 24/7 Availability: Say goodbye to limited learning hours! ChatGPT is available round the clock, providing you with instant access to language support whenever you need it. Whether it's daytime or nighttime, ChatGPT is there to assist you.
  5. Error Correction and Feedback: ChatGPT can help you identify and correct common errors in your writing. By analyzing your sentences and providing constructive feedback, ChatGPT aids in improving your grammar, sentence structure, and overall writing skills.
  6. Multimodal Learning: ChatGPT goes beyond text-based interactions. It can process and respond to audio and visual inputs, making it an immersive tool for practicing listening and comprehension skills. You can even explore English media together with ChatGPT.
  7. Continuous Improvement: As an AI model, ChatGPT is continually updated and refined based on user feedback and advancements in language learning research. This ensures that you receive the most accurate and relevant information for your language development.

Scientifically-published literature on the potential of ChatGPT to be applied as source of information for learning reveal that responses generated by ChatGPT are remarkably good (Hosseini et al., 2023). However, they can also lead to erroneous, misleading, or irrelevant responses (Hosseini et al., 2023)1 and tend to be inconsistent (Elbanna & Armstrong, 2023)2. Picture I attach below describes how it fails in generating consistent information.

Now, what about Bard AI by Google? Here is the response.

We all agree that although possessive pronouns and possessive adjectives exist to state ownership or possession, but those are two different terms. When it is clarified, ChatGPT says that there may be instances where my responses could vary due to some reasons.

  1. Training data: My responses are based on the information available to me up until September 2021. If there have been new developments or changes in the world since then, I may not be aware of them.
  2. Interpretation of queries: Sometimes, the phrasing or context of a question can lead to different interpretations. Depending on how a question is framed, I may provide different answers even if the underlying intent is similar.
  3. Probabilistic nature: As an AI model, I generate responses by calculating probabilities for different sequences of words. In some cases, there may be multiple valid responses, each with a certain probability. This probabilistic nature can result in slight variations in responses.

The responses provided by Google Bard AI regarding inconsistency of responses are somewhat similar to the acknowledgments made by ChatGPT.

While the AI language model, ChatGPT, constantly learns from user interactions and feedback to enhance its responses, it emphasizes the importance for users to critically evaluate the information it provides and cross-reference it with reliable sources. This is particularly crucial for time-sensitive or highly specific topics. By encouraging users to exercise critical thinking and verify information from trustworthy sources, ChatGPT aims to promote responsible and informed use of its generated content.

  1. Hosseini, M., Rasmussen, L. M., & Resnik, D. B. (2023). Using AI to write scholarly publications. Accountability in Research, 1–9. 10.1080/08989621.2023.2168535 ↩︎

  2. Elbanna, S., & Armstrong, L. (2023). Exploring the integration of ChatGPT in education: Adapting for the future. Management & Sustainability: An Arab Review, ahead-of-print. 10.1108/MSAR-03-2023-0016 ↩︎