ChatGPT has been dominating the internet since its debut in November 2022. There isn’t much that the artificial intelligence (AI) and machine learning (ML) powered bot cannot achieve, including writing complex codes and telling complicated stories. In an effort to facilitate automation, numerous large companies, like Microsoft, have tried to integrate the bot on their websites. However, this raises the straightforward question: Is there anything ChatGPT cannot do? Read below to know further.
What is ChatGPT and how does it work?
Elon Musk, Sam Altman, John Schulman, and other prominent figures in the IT industry helped create the US-based OpenAI, an AI research facility. A human-like text answer to inquiries is provided by ChatGPT, which was developed using the company’s proprietary GPT-3 big language model (probability distribution over a series of words). Currently available as a prototype, ChatGPT is available for free. However, according to certain media sources, OpenAI may eventually monetize access to the programme.
ChatGPT requires training just like any other AI-powered tool in order to produce the necessary responses. For this, researchers applied the supervised learning and reinforcement learning machine learning models. OpenAI asserted on its official website that ChatGPT “sometimes writes plausible-sounding but incorrect or nonsensical answers.” It will take advanced problem-solving in fields like reinforcement learning and others to resolve that difficulty.
Since ChatGPT’s knowledge cut-off date has been set to 2021 only, it is also unable to respond to any questions about events that occurred after that year.
What ChatGPT cannot do?
While it is possible to argue that the aforementioned restrictions may be overcome with more AI training and rewriting of the program’s code to better hone it, there is still one really straightforward duty that any AI cannot perform—creating.
The easiest way to describe this is through Tharoor’s customary caustic tweet. Nishanth Vijayan (@nishanth who), a Twitter user, shared a snapshot of a ChatGPT chat on January 15 in which he requested that the programme create an absence notice “in the manner of Shashi Tharoor.”
The well-known MP and word-smith Shashi Tharoor wrote in response, “Hilarious. But I really can’t see anything so jejune!” For those unaware (and understandably so), “jejune” means naive, simplistic, and superficial.
You’re not alone if you notice a striking resemblance between Tharoor’s statement and OpenAI’s justification of ChatGPT’s limitations. We both agree that we asked ChatGPT to accomplish the unthinkable. But on the other hand, because it possesses the capacity for inventiveness, human prowess will ultimately prevail in this situation.
While that may seem intellectual, innovation is ultimately driven by imagination. As a result, ChatGPT won’t ever be able to innovate and will always require human input to do the extra step of fine-tuning.
Keep watching our YouTube Channel ‘DNP INDIA’. Also, please subscribe and follow us on FACEBOOK, INSTAGRAM, and TWITTER