Community Knowledge Synthesis in Large Language Models (LLMs)

The community aspect of the LLMs must be remembered

The creation and widespread adoption of ChatGPT and other Large Language Models (LLMs) represent a remarkable achievement for the global community. These powerful AI systems are revolutionizing the way we generate text and engage in conversations, offering tremendous potential for various industries to leverage machine intelligence. While LLMs are not flawless and occasionally produce imperfect or inaccurate outputs, they serve as invaluable tools for sparking ideas and generating content from scratch. These models are trained and synthesized using the vast amount of freely available information on the Internet, and recent iterations like Bing Chat, Google's Bard, and ChatGPT Plus Plugins can even connect to the internet to stay up to date. However, concerns have been raised by researchers and content creators regarding the use of original works without proper credit and compensation. While these concerns are valid, LLMs also highlight the intricate relationship between existing knowledge and the creation of new content, demonstrating an intermediate process.

LLMs are characterized by an extensive array of mathematical functions derived from existing knowledge obtained from the internet and other sources. This wealth of knowledge has been amassed by human society over its entire existence, initially preserved in printed form, and now in digital media and on the internet. The advent of the internet has partially democratized collective human knowledge, making information more accessible. LLMs and their various iterations now take this accessibility to the next level by presenting synthesized versions of collective knowledge. These two revolutions represent significant strides in making knowledge and information more readily available to the masses.

LLM as the Knowledge Facilitator

To illustrate further, the internet has opened up the once-gated realm of information, making it easily accessible and affordable to a large portion of the global population. With a mobile phone and an internet connection, people can now visit websites, read any content they desire, and even share their ideas freely. This newfound accessibility has created new opportunities, for example, enabling individuals from suburban areas or villages in countries like India or Indonesia to pursue PhDs in the USA and engage in cutting-edge research. Even two decades earlier, such opportunities were confined to the upper class of society or strictly to the urban population. Moreover, these individuals can use the same internet to connect with their peers, professors, and even school teachers back in their home countries. This seamless and continuous discourse has flourished over the past decade, courtesy of the internet. The internet has been a great force in partly reducing the vast gaps in knowledge accessibility between different classes in our society.

While the internet is undeniably valuable, it is also huge and unfiltered. The synthesized knowledge presented by LLMs will play a crucial role in facilitating greater access to the human knowledge database, particularly for underrepresented and marginalized communities. A conversational LLM like ChatGPT can be viewed as a friendly teacher or guide, helping individuals navigate the overwhelming wealth of raw information in the human knowledge database. The Khan Academy has started using conversational LLMs to bring a personal teacher for everyone. More such efforts are underway.

The Uniqueness of Innovation

Furthermore, LLMs have the ability to unearth new insights and generate novel ideas. The novelty aspect of generative AI showcases that innovation and content creation are not exclusive to a select few with unique human skills. In reality, innovation heavily relies on existing knowledge. This is not to discount the contribution of human ingenuity to the fascinating novelties of the modern world. However, artificial intelligence engines are democratizing access to raw information, empowering a greater number of individuals to engage in innovation.

Final Bits: The Collective Effort behind LLMs

Ultimately, the raw material for artificial intelligence engines is the collective human knowledge database, which is owned by society as a whole. Therefore, the synthesis of knowledge by LLMs should be perceived as a collective endeavor. While companies like Google, OpenAI, and Meta play prominent roles in the AI revolution, we, as individuals, have also contributed and are still contributing to this transformation, even if in smaller capacities. The LLMs were not possible if information and knowledge-sharing were guarded for the elite few. Therefore, this momentous achievement is a cause for celebration for the human community as a whole, barring the division of the economic classes.