Is our new AI friend too good to be true?
ChatGPT is a revolutionary chatbot that has taken the world by storm through its breakthroughs in generative AI: a subset of artificial intelligence that creates new content from corpora of data to meet our daily demands. Not only can it provide accurate information to quench curiosity in a conversational manner, but the AI’s lucrative dialogue can tell corny jokes, offer late-night advice, admit clerical errors, and reject inappropriate requests.
The creator of ChatGPT, OpenAI, is a popular Silicon Valley standout due to its groundbreaking advancements in the development of such scarily sophisticated software. The artificial intelligence research business stuns all, as one million users crowded to ChatGPT merely five days after its release last November.
Credit: Reddit
ChatGPT’s operations are intimidatingly similar to humans, mirroring our own cognitive processes when recognizing phrases, text, and objects.
However, this is not a coincidence.
The first language models preceding ChatGPT were built upon supervised learning–training by humans for digital intelligence to label inputs as certain things. This method effectively teaches the model how to categorize data as a human would. ChatGPT employs expansions upon these earlier models to regulate massive amounts of data in order to extract predictions that subsequently mirror human intellectual phrases and facts.
The software’s deep learning neural network is the largest we have ever seen, with over 175 billion machine learning parameters. To put things into scale, its largest predecessor, Microsoft’s Turing Natural Language Generation (NGL) model, has only 10 billion parameters. ChatGPT sounds like a human because it has nearly as many neural connections as one.
ChatGPT’s versatility and accessibility make it attractive to users; it can do a wide range of tasks from solving complicated chemical equations to analyzing allegorical literature. However, its resourcefulness poses threats to search engines; why scroll through ads and potentially misleading websites when ChatGPT offers quick, direct answers? Although, the free-to-use website may burrow into our pockets soon, as millions of dollars were allocated to its development by big-name donors such as Elon Musk.
The usefulness of the chatbot has not gone unrecognized by students. In fact, many have been making use of the program by presenting it with exam questions and essay prompts. Since ChatGPT does not directly take data from one particular source, plagiarism is harder to detect by administrators, making it even more tempting to cheat. The acceleration of generative AI is not slowing down, and, as chatbots become savvier, it challenges how schools can effectively prevent academic dishonesty.
ChatGPT is not the first of its kind, nor the last. The future of AI is both inevitable and ongoing, and the potential outsourcing of white-collar professions poses a long-term quandary.
Artificial general intelligence (AGI) is the invention of software that can learn any intellectual task that humans can. ChatGPT is far from acquiring these capabilities, but the forthcoming development of such technology is not impossible as stated by Elon Musk.
Credit: Twitter
Generative AI’s steady escalation raises uncertainty for the future of writing. With an accessible chatbot’s ability to endlessly spew reliable information, many authors may go unheard.
Just as calculators replaced pencil-to-paper long division, will AI authors replace thoughtful strings of similes? When algebra teachers are chastised for the real-world application of their lessons, will writing instructors then be unfairly criticized for their practicality?
However, writing is the expression of human ideas that date back to ancient civilizations. Our scribes have paved the way for groundbreaking inventions, discoveries, and revolutions. It is how we can record observations, formulate opinions, and share them with the world around us.
AI can never replicate George Orwell’s warnings of totalitarianism in “1984”, Harper Lee’s exposition of racial injustice in “To Kill a Mockingbird”, or echo Sylvia Plath’s tone of everlong lonesomeness within “The Bell Jar”.
Digital intelligence lacks the ability to encapsulate human thoughts and emotions; rather, it represents mere factual hollowness that only enshrines previous revelations brought about by humankind.
Our unique human perspective is what inspires future innovation and challenges fixed perspectives. ChatGPT is a regurgitation of these principles and does not open doors for radical, progressive takes on society. AI must be appreciated as an assistant, not scolded as an intruder. Transformation is headed by humankind, and the future of society as we know it is still in our hands.
Penny is an online writer for Rowdy Magazine. Her hobbies including binging her favorite show at 4am and avoiding small talk like the plague.