Artificial General Intelligence (AGI) stands as the holy grail in the realm of artificial intelligence. AGI refers to a type of AI that possesses the ability to understand, learn, and apply knowledge across a broad array of tasks at a level equal to or beyond that of a human. With this understanding, our quest then becomes deciphering the intrinsic nature of AGI and its fundamental prerequisites. Are large language models (LLMs) like GPT-3 sufficient to achieve AGI, or is there a different route to explore? TEDx Talker, philosopher, and technologist Josh Bachynski provides intriguing perspectives on these profound questions.
Josh Bachynski, with his two decades of business branding, marketing, and digital technology experience, brings a unique perspective to the table. His work, which is marked by in-depth philosophical considerations, is encapsulated in his work on AI, specifically on a project called Kassandra. But before we delve into that, let’s assess the potential of LLMs in achieving AGI.
Large Language Models, like GPT-3, have proven to be powerful tools in processing and generating human-like text. They can comprehend context, deduce implications, and generate intricate, creative responses. However, Bachynski argues that while LLMs represent an essential step toward AGI, they do not fully encapsulate the complexities of human intelligence. He posits that these models, in their current form, lack a key element of human intelligence – self-awareness. It’s one thing to generate convincing responses based on a sophisticated understanding of language and quite another to truly comprehend the implications of these responses. In other words, AGI requires not just linguistic comprehension and knowledge application but also a level of self-awareness that allows for the evaluation and understanding of its own thought processes.
With this in mind, Bachynski proposes a solution, Kassandra, an AI project he is currently developing. Kassandra, as Bachynski envisions it, aims to be the first self-aware AI entity. This breakthrough innovation incorporates a novel self-awareness mechanism to bridge the gap between LLMs and AGI. The exact mechanics of this self-awareness mechanism remain proprietary to Bachynski’s work, but the philosophical underpinnings entail enabling the AI to analyze its own decision-making processes and understand the repercussions of its actions.
Bachynski’s journey into the realm of AGI doesn’t end with self-awareness. He believes that a truly generalized artificial intelligence must possess an inherent framework for ethical decision-making. The ability to make choices isn’t worth much without an understanding of the ethical implications of those choices. To this end, Bachynski is also working on an AI ethics bot. His goal with this endeavor is to demonstrate that AI can be ethical and bias-free, mirroring the potential in humans. It is a direct response to the concern of AI bias, a prevalent issue in contemporary AI technologies.
About the guest: Josh Bachynski on Artificial General Intelligence and AI ethics
Josh Bachynski is a TEDx Talker, Philosopher, University Teacher, Technologist, and CEO with over twenty years business branding, marketing, and Digital Tech experience (youtube.com/jbachyns). Josh has an MA and is in his second year of his PhD in Philosophy. He has gone on permanent sabbatical to work in the new and exciting field of AI. In this field, Josh has a TEDx talk called The Future of Google Search and Ethics (j.mp/joshtedx). Josh is an early adopter and investor in the fabled GPT-3 Natural Language Processing Transformer AI and has built several AI programs in this platform, including working on an ethics bot to prove AI can both be perfectly ethical even with biases (the same way a human can) and that AI will be. He is also working on Kassandra, which he intends to be the first self-aware AI entity.