Название: The Creation of a Conscious Machine: The Quest for Artificial Intelligence, 2nd Edition
Автор: Jean E. Tardy
Издательство: Mercury Learning and Information
Серия: MLI Generative AI Series
Год: 2024
Страниц: 327
Язык: английский
Формат: pdf, epub (true)
Размер: 10.1 MB

This book presents a groundbreaking journey into the world of Generative AI technology and offers an in-depth look at the prospect of AI achieving consciousness. The book navigates through various historical and modern perspectives on AI, from ancient myths to the Turing Test to the latest in technological advancements. It covers the theoretical and practical aspects of creating a conscious AI, including the specifications for synthetic consciousness and the integration of AI with human cognition. The book questions whether generative AI can meet the traditional criteria of consciousness and how this might be realized.

In 2011, I published the first version of The Creation of a Conscious Machine, tracing the evolution of the quest for AI from antiquity to the present, culminating in a definition of artificial consciousness expressed in terms of clearly defined and achievable specifications. In 2015, I published The Meca Sapiens Blueprint, a complete system architecture to implement consciousness in autonomous agents.

Recently, a new type of AI system, based on Generative AI technology, was officially released. These systems are very different from autonomous agents. They are deep learning systems that integrate large amounts of existing documentation to produce conversational output that can be stunningly convincing. This raises a question: Can generative AI meet the Meca Sapiens’ specifications of consciousness, and, if so, how?

In this new, expanded, version of The Creation of a Conscious Machine, we describe how the core conditions of consciousness, initially intended for synthetic agents, can also be applied to generative AI, and we outline how this can be achieved. The original ­content is also updated and clarified in light of recent ­developments in AI.

In late 2022, OpenAI publicly released its first version of ChatGPT [OPENAI22], a generative AI chatbot. It became an overnight sensation. Countless individuals were amazed by its output, some even claiming the program was sentient! ChatGPT was soon followed by similar programs such as Bard from Google [BARD23] and Copilot by Microsoft [CPLT23]. A few months later, about a dozen competing generative AI applications were available, some running on independent platforms and others providing added functionality on top of existing ones.

Generative AI is thus very different from conventional robotic AI. In robotic or agent types of AI, the human individual interacts with an entirely synthetic entity whose “intelligence” is generated by computer logic. With generative AI, the individual user interacts with something else: an integrated version of the shared cogitations of thousands of humans. The GenAI apps are like portals that allow individual humans to communicate directly with a portion of the collective mind of their species. However, these apps are also only portals since they don’t internally generate semantic models of their environment or themselves. The intelligent output they produce is entirely dependent on the training set drawn from the corpus and whose content is (mainly) generated by human beings. Without this human production, the apps would not exhibit any intelligence whatsoever.

Features:
Specifies the design choices and implementation strategies that must be followed to successfully build machines that are conscious
Explores the entire spectrum of AI development, from ancient origins to the potential future of conscious machines
Offers a critical examination of the Turing Test, its variations, and its relevance to modern AI
Provides insights into the potential paths and challenges in achieving synthetic consciousness in AI




ОТСУТСТВУЕТ ССЫЛКА/ НЕ РАБОЧАЯ ССЫЛКА ЕСТЬ РЕШЕНИЕ, ПИШИМ СЮДА!