By: Joshua Xu
Technology’s integration into our lives has become perpetually more prolific, with the merits and implications of Artificial Intelligence remaining a central locus of discussion. While AI has already realized significant impacts in industry, the introduction of ChatGPT has made this technology accessible to the general public, making discourse surrounding AI strikingly more relevant.
What is ChatGPT?
ChatGPT is an Artificial Intelligence (AI) chatbot developed by OpenAI, a San Francisco-based startup that was released in November of 2022 (Hughes). OpenAI was co-founded in 2015 by Elon Musk and Sam Altman and is backed by numerous investors, including Microsoft, with its stated mission being to ensure AI benefits “all of humanity”.
ChatGPT is one of many examples of generative AI. While any algorithm can generate output, generative AI is unique in its focus on matching patterns in data to create convincing content. Other examples of generative AI models include Dall-E, which generates digital images based on natural descriptions, and Lensa, which specializes in AI-created avatars (Engler).
While such forms of AI have existed prior to ChatGPT’s release, it is the ability to create human-like responses that makes ChatGPT unique. The chatbot is powered by a large language model (LLM): deep learning algorithms that can recognize, synthesize, and generate text from training based on massive datasets. ChatGPT uses GPT-3.5, an upgrade of GPT 3 and one of the largest language processing AI models, although OpenAI released GPT 4 on March 14 - which now allows ChatGPT to handle image inputs among other improvements - to GPT-Plus members for $20 per month (Browne). Both the user-interface and interacting with the chatbot are simple: by typing a question or message, users may write or debug code, explain concepts, excerpt data from texts, and tell stories.
What makes ChatGPT so popular?
Primarily, ChatGPT’s free cost via OpenAI’s website makes its various possible applications highly accessible, resulting in it reaching over 100 million users after 2 months and becoming the fastest-growing consumer application in history (Browne). Unlike search engines like Google, ChatGPT is conversational - providing human-esque responses and dialogue that provides the convenience of allowing the user to make direct demands through an easily comprehensible interface.
How effective is ChatGPT?
While the sudden success of ChatGPT appears novel, on a technical level, it is not functionally different from previous AI systems - it is just better at similar tasks. ChatGPT represents a transition, demonstrating that applying AI to creative and expressive jobs opens a realm of new possibilities as opposed to traditionally repetitive tasks.
ChatGPT has the extraordinary ability to perform writing with more significant implications than initially apparent. For example, the ability to participate in back-and-forth conversations with the AI highlights the possibility of human-machine hybrid work - humans can guide AIs and correct mistakes. The limits of the language AI model are unknown and offer still undiscovered possibilities.
However, being a complex algorithm generating meaningful text, the AI itself cannot recognize what it does not know because they are not trained to objectively evaluate the accuracy of its outputs. In a sense, the AI can not explain what it does or how it does it, leaving the possibility for biases and unethical action (Browne). Other times, the AI can simply fail to provide an answer to inquiries that deviate from the norm.
Impact of Generative AI on Education
Evidently, the proliferation of ChatGPT can make it difficult to credibly assess the quality of student learning. However, such challenges are not new, but rather exacerbated by the introduction of generative AI. In academia, the immediate reaction might be to adopt more traditional assessment strategies, rejecting the integration of generative AI through preventative measures to deter students from using similar tools. However, it is clear that ChatGPT and similar technologies will continue to shape industry and, as such, teaching and assessment practices will have to be revised in a way that allows students to adopt these technologies (Sinha). For example, by contextualizing project-based assessments to localized contexts, students may conduct an exploration of ChatGPT and related tools but ensure an understanding of higher-order elements that require synthesis and evaluation. In response to the growing influence of generative AI in the classroom, Columbia University’s Jin Kuwata states that “I see these kinds of tools as not just things that complete tasks, but entities we learn to work with in partnership to enrich our understanding of the world” (Gilbard). Tools like ChatGPT may push teachers and educational institutions to design increasingly more active, culturally relevant learning environments in order to equip students to the perpetual changes technology brings to industry and beyond.
Browne, Ryan. “All You Need to Know about Chatgpt, the A.I. Chatbot That's Got the World Talking and Tech Giants Clashing.” CNBC, CNBC, 8 Feb. 2023, www.cnbc.com/2023/02/08/what-is-chatgpt-viral-ai-chatbot-at-heart-of-microsoft-google-fight.html.
Engler, Alex. “Early Thoughts on Regulating Generative AI like Chatgpt.” Brookings, Brookings, 28 Feb. 2023, www.brookings.edu/blog/techtank/2023/02/21/early-thoughts-on-regulating-generative-ai-like-chatgpt/.
Hughes, Alex. “CHATGPT: Everything You Need to Know about OpenAI's GPT-4 Tool.” ChatGPT: Everything You Need to Know about OpenAI's GPT-4 Upgrade | BBC Science Focus Magazine, BBC Science Focus Magazine, 16 Mar. 2023, www.sciencefocus.com/future-technology/gpt-3/. Sinha, Saurabh, et al. “How CHATGPT Could Revolutionize Academia.” IEEE Spectrum, IEEE Spectrum, 9 Mar. 2023, spectrum.ieee.org/how-chatgpt-could-revolutionize-academia.