More than one million users in five days, top trending on social media, part of dinner table talks, debates on academic and practitioners’ forums and hundreds of news articles and blogs in less than a month, has made ChatGTP one of the most disruptive technologies of recent times. But why? What makes ChatGTP more special than other LLMs? Is it sentient? Is it a hint towards technology singularity? We have penned our observations and present the topic for debate.
An AI system is sentient when it can perceive and feel as humans, possess consciousness and be self-aware. ChatGPT in some instances is said to be sentient as it exhibits features of anthropomorphism (attribution of human traits) by answering follow-up questions, demonstrating complex scenarios with examples, and notifying users about inappropriate requests. On a technical level, ChatGPT is a non-sentient LLM that synthesizes information by mimicking human language.
ChatGPT’s unanticipated popularity could also be attributed towards super intelligent AI required for technological singularity. Technological singularity is a future event where AI will be capable of recursive self-improvement and ever-increasing AI capabilities will be beyond human comprehension. We are intrigued how an LLM could hint towards singularity. To answer that we compare two LLMs.
ChatGPT is a variant of GPT-3 – another LLM- but the difference between both is the size and type of data used for training. The major difference between GPT-3 and ChatGTP is in training model of both chatbots. GPT-3 is trained on large text data including books, articles and websites. Whereas ChatGPT is trained primarily on conversational data by including humans in the loop. Due to the conversational design, ChatGTP exhibits human-like features. While GPT-3 despite having high volumes of data, seem less interactive and rather provide straightforward summaries for the queries. Also, GPT-3 works with 175 billion parameters that require extensive computational resources and is only suitable for high-power devices. On the other hand, ChatGPT only requires 117 million parameters, therefore it is lightweight, device friendly and highly accessible.
We agree that ChatGPT is non-sentient, but we are sceptical about the conversational and argument building attribute of ChatGTP as this could lead to self-learning AI. To create a super-intelligent AI, sentience is not a necessary condition. This leads us to believe that singularity of technology might be edging closer, but there is more work to be done – as one can see when we ask ChatGPT to help us write an article on “whether it is bringing us closer to sentient AI or singularity”.
Have a look how ChatGPT responded to the same statement!
ChatGPT is a state-of-the-art language model developed by OpenAI that utilizes deep learning techniques to generate human-like text. It is designed to be highly proficient in tasks such as language translation, text summarization, and question answering. However, it is important to note that ChatGPT is not sentient and does not possess consciousness or self-awareness.
The term "sentient AI" refers to a type of artificial intelligence that has the ability to perceive and respond to its environment in a manner similar to that of a living organism. Sentience is often associated with consciousness and self-awareness, but the two are not necessarily synonymous. Currently, no AI system has been developed that can truly be considered sentient.
The Singularity is a term that refers to a hypothetical future point in time when artificial intelligence will have surpassed human intelligence, leading to an acceleration in technological progress and a profound impact on humanity. This point is often associated with the emergence of sentient AI, but it is important to note that the Singularity is a highly speculative concept and its likelihood of occurring is the subject of much debate among experts in the field.
While ChatGPT is a highly advanced AI system, it is not sentient and does not bring us closer to the Singularity. It is a tool that can be used to generate human-like text and perform a variety of language-based tasks, but it does not possess consciousness or self-awareness. Additionally, its capabilities are limited to the specific tasks it was trained on and it cannot perform tasks that are outside of its programmed abilities.
It is also important to note that the Singularity, if it were to occur, would have far-reaching implications for humanity and society. It would likely lead to rapid technological advancements and changes in the way we live and work. However, it is also important to consider the potential negative consequences, such as job displacement and ethical concerns.
In conclusion, while ChatGPT is a powerful language model, it is not sentient and does not bring us closer to the Singularity. Sentience and the Singularity are complex and highly speculative concepts that are the subject of ongoing research and debate in the field of artificial intelligence. As with any new technology, it is important to consider both the potential benefits and drawbacks and to approach its development and deployment with caution.
Authors:
Professor Kok-Leong Ong
(Director, Enterprise AI and Data Analytics Hub - School of Accounting, Information Systems and Supply Chain, RMIT College of Business and Law)
Dr Samar Fatima (Enterprise AI and Data Analytics Hub - School of Accounting, Information Systems and Supply Chain)
