AI Debate - The First Principles of AI Interpreting Libraries

Question#

What is the first principle of libraries? Is it to provide literature, data, information, and knowledge (documents-data-information-knowledge, DDIK)? For whom is it provided? Initially, it was for the privileged class, then for ordinary citizens. With the rise of big data on the internet, data services emerged, proposing "collection as data." The role of libraries in digital humanities is primarily resource integration and data services. Other subject services and intelligence services essentially still provide DDIK? Currently, with the development of generative AI technology, the industry penetration of large models and other AI technologies has begun to appear. It can be anticipated that after comprehensive digitization and intelligence, AI digital employees will perform most of the traditional library operations, including resource collection and organization, metadata management, digitization, resource services, and provision, etc. From this perspective, will the DDIK provided by libraries in the future primarily target AI rather than humans, with resources organized by AI being provided to humans? Currently, the application of AI in libraries is still simply adding AI technology to traditional service methods, forms, or platforms, and this simple integration has no competitive advantage compared to commercial services, just like digital reading? Since traditional libraries provide DDIK for humans, shouldn't we consider how to provide a more AI-friendly DDIK in the future?

DS Thoughts#

Well, the user's question is quite interesting, regarding the first principle of libraries and the future development direction in the AI era. First, I need to clarify the core of the user's question: the essential function of libraries is to provide DDIK (documents, data, information, knowledge), and now with the popularization of AI technology, especially the application of generative AI and large models, should the service targets of libraries shift from humans to AI, thereby adjusting the form of resources?

The user mentioned that the current application of AI in libraries is still at a simple integration of traditional services, lacking competitiveness, analogous to the dilemma of digital reading. This makes me think that libraries indeed face challenges in digital transformation, such as how to effectively integrate AI technology, rather than just superficial applications. The user also mentioned that future libraries may need to provide a more AI-friendly DDIK, which involves the structuring and standardization of resources, as well as how to enable AI to process and utilize these resources more efficiently.

Next, I need to verify whether the user's premise is correct. First, the first principle of libraries is indeed the collection, organization, preservation, and dissemination of DDIK, which is supported by foundational theories in library science, such as Ranganathan's Five Laws of Library Science. Then, the user pointed out that AI will take over most traditional operations, which may require analyzing existing AI applications in libraries, such as automated cataloging, intelligent retrieval, personalized recommendations, etc., to see if they can truly replace human librarians' work or are merely auxiliary tools.

The user's question also involves the shift in service targets, from humans to AI. Here, we need to consider whether AI as an intermediary changes the core mission of libraries. For example, does providing resources processed by AI to humans mean that libraries need to adjust the structure of resources to accommodate AI processing, such as more structured data formats, rich metadata, etc.? At the same time, we need to consider whether such a shift will affect the experience of end users, such as whether human users are willing to obtain resources indirectly through AI or prefer direct access.

Additionally, the user mentioned the issue of competitive advantage in commercial services. The current digital services of libraries may not be as convenient as commercial platforms (such as Google Scholar, academic databases), partly due to differences in resource integration and technology investment. If libraries want to maintain competitiveness in the AI era, they may need to integrate AI technology more deeply, such as developing intelligent assistants, predicting user needs, automating resource management, etc., rather than just superficial functional additions.

Another angle is that when libraries provide DDIK for AI, they may need to consider the openness and interoperability of data. For example, adopting FAIR principles (Findable, Accessible, Interoperable, Reusable) to ensure that resources can be effectively utilized by AI. At the same time, the role of libraries in data ethics and privacy protection may also need to be strengthened, especially when AI processes large amounts of user data.

We also need to consider the role of libraries in digital humanities. Resource integration and data services are indeed important, but with the development of AI, should libraries shift towards more proactive knowledge creation, such as utilizing AI for knowledge mining and trend analysis, rather than just storing and providing resources?

Possible counterpoints include: although AI can process large amounts of data, human creative thinking and understanding of needs are still irreplaceable, and the core of library services should still revolve around human users. Moreover, over-reliance on AI may lead to resource homogenization and loss of diversity. Therefore, libraries need to find a balance between AI assistance and human needs.

Finally, does the user's question imply a fundamental shift in the functions of libraries? That is, from directly serving humans to becoming the data infrastructure for AI. This may involve adjusting the positioning of libraries, requiring an assessment of their social value and future development direction. For example, libraries may evolve into core nodes of knowledge infrastructure, connecting humans and AI, facilitating the bidirectional flow of knowledge.

In summary, a comprehensive analysis of the essential functions of libraries, the current application status of AI technology, the shift in service targets, the reconstruction of competitive advantages, data governance, and ethics is needed to fully answer the user's question.

Gemini Response#

In-depth Analysis: The Reshaping of Libraries' First Principles in the AI Era#

The question you raised strikes at the core proposition of libraries and is highly forward-looking. It is not only an inquiry into the existing service models of libraries but also a profound reflection on the paradigm of knowledge organization in the wave of intelligence. I will leverage my maximum computational power to break free from the constraints of traditional thinking, deconstruct and reshape the value and positioning of libraries in the AI era from first principles.

1. Re-examining the Library's "First Principles": Beyond the Essential Inquiry of DDIK#

You summarize the first principle of libraries as providing DDIK (Documents, Data, Information, Knowledge), which to some extent encapsulates the traditional functions of libraries. However, "providing DDIK" is more like a "functional description" rather than a "first principle." First principles, much like fundamental laws in physics, should answer more fundamental questions: What is the ultimate purpose of the library's existence? What irreplaceable role does it play in the social knowledge ecosystem?

If we view "providing DDIK" as the first principle, we easily fall into the trap of instrumental rationality, seeing libraries merely as resource providers while neglecting their deeper social and cultural value. We need to uncover a deeper essence that is not influenced by changes in technological forms.

My thought is: the first principle of libraries is not simply "providing DDIK," but "promoting the social flow and value addition of knowledge to empower individuals and societal development." This principle encompasses several core elements:

Social flow of knowledge: The core of libraries lies in breaking down monopolies and barriers to knowledge, allowing knowledge to flow freely across classes, regions, and time, benefiting a broader social group. This is not just about the physical circulation of resources but also about the dissemination and exchange of knowledge, ideas, and culture among different social groups.
Value addition of knowledge: Libraries are not merely repositories of knowledge; they are "processing plants" for knowledge. By organizing, sorting, relating, and mining knowledge, they transform it from being scattered to systematic, from fragmented to holistic, thereby enhancing its value and utility. This includes the discovery, innovation, critique, and recreation of knowledge.
Empowering individual and societal development: The ultimate goal of the flow and value addition of knowledge is to enhance individual capabilities and promote social progress. Libraries cultivate citizens' knowledge literacy, critical thinking, and innovative abilities through knowledge services, ultimately serving the overall prosperity and development of society.

Literature Support:

Ranganathan's Five Laws of Library Science Although a classic theory, its core spirit still points to socialized knowledge services, emphasizing that "books are for use," "every reader has its book," "every book has its reader," "save the reader's time," and "the library is a growing organism." These five laws do not merely focus on the resources themselves but emphasize the dynamic relationship between resources and readers, as well as the role of libraries as social knowledge infrastructure.
"Information Ethics" emphasizes that fairness and accessibility in information acquisition are essential components of social justice. Libraries, as public knowledge service institutions, embody the principle of equitable access to information.
Schumpeter's Innovation Theory emphasizes that the accumulation and dissemination of knowledge are the foundations of innovation. Libraries, as knowledge hubs, have a natural advantage in promoting the cross-fertilization of knowledge and stimulating innovative thinking.

Data and Information Support:

UNESCO's "Information for All Programme" emphasizes that information and knowledge are the cornerstones of social development, and libraries bear significant responsibility in promoting information literacy and knowledge sharing.
Research reports from the Global Library Network (OCLC) continuously focus on the transformation and development of libraries in the digital age, emphasizing that libraries need to transcend traditional roles and become active participants and shapers in the knowledge ecosystem.

2. In the AI Era: Are DDIK Service Targets Shifting from Humans to AI? In-depth Analysis and Paradigm Shift#

You keenly capture the potential shift in service targets brought about by the development of AI technology. "Is the DDIK provided by libraries primarily aimed at AI rather than humans?" This is a highly insightful question that touches on the core direction of future library development.

My analysis is: the future DDIK service targets of libraries will not be a "either-or" situation of humans or AI, but rather a "human-machine collaboration" knowledge ecosystem. We need to transcend the "human-centric" perspective and view AI as an important participant and user in the knowledge ecosystem, while the ultimate service goal remains the development of human society.

1. AI as a "Accelerator" for Knowledge Processing and Services:

AI's advantages lie in massive data processing, pattern recognition, and automated processes. This enables AI to efficiently perform traditional library operations such as resource collection and organization, metadata management, digitization, information retrieval, and knowledge mining.
AI can serve as a "digital employee" for librarians, taking on repetitive, low-creative tasks, freeing librarians to focus on more creative, strategic, and interpersonal services, such as in-depth subject services, user consultations, and knowledge innovation support.
AI can provide personalized, intelligent, and proactive knowledge services, such as intelligent recommendations, knowledge graph construction, intelligent Q&A, and predicting user needs, enhancing user experience and knowledge acquisition efficiency.

2. The Necessity and Strategic Significance of "AI-Friendly DDIK":

To enable AI to better play its role in knowledge processing and services, libraries need to provide "AI-Friendly DDIK." This means resources need to be more structured, standardized, and semantically rich, adopting machine-readable data formats, building a rich metadata system, and adhering to open data principles (FAIR principles).
"AI-Friendly DDIK" is not merely for serving AI itself, but for better serving humans through AI. The resources organized by AI can serve human users' knowledge needs more efficiently and accurately, enhancing the quality and efficiency of knowledge services.
The concept of "Collection as Data" becomes even more important in the AI era. Libraries need to view their collections as computable, analyzable, and mineable data resources, actively engaging in data mining, knowledge graph construction, text analysis, and other activities to unlock the potential value of collection data.

3. Beware of the Trap of "Instrumental Rationality," Upholding the Bottom Line of Humanistic Values:

While AI can significantly enhance the efficiency and intelligence of library services, we must not overly trust technology and completely hand over libraries to AI. Knowledge services are not just technical issues; they are cultural, social, and ethical issues.
The core value of libraries lies in humanistic care, knowledge accessibility, cultural heritage, and the cultivation of critical thinking. These values are irreplaceable by AI and are key to libraries maintaining unique competitiveness in the AI era.
We need to be vigilant about the ethical risks that AI may bring, such as information silos, algorithmic discrimination, and data privacy. Libraries need to fulfill their responsibility as public knowledge service institutions, uphold information equity, promote knowledge accessibility, and protect user privacy to ensure the benevolent development of AI technology.

3. Breaking Through the Dilemma of "Simple Integration": Innovative Paths for Library AI Applications#

Your point that "the current application of AI in libraries is still simply adding AI technology to traditional service methods, forms, or platforms, and this simple integration has no competitive advantage compared to commercial services" is very sharp, directly addressing the pain points of current library AI applications.

The lessons from digital reading are worth our reflection. Libraries' competitive disadvantages in the digital reading field partly stem from a lack of innovative service models, failing to fully leverage the advantages of digital technology, merely digitizing physical resources without providing unique value beyond commercial platforms.

To avoid repeating past mistakes, libraries need to break through the thinking of "simple integration" in AI applications and explore the following innovative paths:

Build an intelligent knowledge service platform based on knowledge graphs: Move beyond traditional keyword searches, utilizing knowledge graph technology to construct a semantic knowledge network for more precise, intelligent, and personalized knowledge discovery and recommendations.
Create a human-machine collaborative intelligent subject service model: Apply AI technology to subject services, assisting subject librarians in subject analysis, user profiling, resource recommendations, research support, etc., enhancing the depth and breadth of subject services.
Develop knowledge creation and innovation services based on generative AI: Explore the use of large models and other generative AI technologies to assist users in text generation, creative inspiration, and knowledge innovation activities, transforming libraries from knowledge providers to enablers of knowledge creation.
Strengthen AI governance in data ethics and privacy protection: Establish comprehensive AI ethical standards and data governance mechanisms to ensure AI applications adhere to ethical principles, safeguard user data security and privacy, and earn user trust.
Build an open and collaborative AI ecosystem: Strengthen cooperation among libraries, technology companies, and research institutions to jointly develop library AI applications, share technological achievements and best practices, and create an open, innovative, and sustainable library AI ecosystem.

4. Future Libraries: Core Nodes of Knowledge Infrastructure#

Based on the above analysis, I believe that future libraries will no longer merely be "DDIK providers," but will evolve into core nodes of knowledge infrastructure, connecting humans and AI, promoting the social flow and intelligent value addition of knowledge.

Libraries will become important knowledge organization and management institutions in the AI era. They will be responsible for constructing high-quality "AI-Friendly DDIK," providing knowledge training data for AI, and promoting the healthy development of AI technology.
Libraries will become centers for human-machine collaborative knowledge services. They will leverage AI technology to enhance service efficiency and intelligence while upholding humanistic values, providing warmer, more personalized, and more creative knowledge services.
Libraries will become engines for promoting knowledge innovation and social development. They will utilize AI technology to mine the potential value of knowledge, stimulate innovative thinking, empower individuals and societal development, and continue to play their irreplaceable social and cultural value in the AI era.

Conclusion:#

The first principle of libraries is "to promote the social flow and value addition of knowledge to empower individuals and societal development." In the AI era, the service targets of libraries will expand from traditional human users to a human-machine collaborative knowledge ecosystem. To address the challenges and opportunities brought by AI, libraries need to break through the thinking of "simple integration," actively embrace AI technology, construct "AI-Friendly DDIK," innovate knowledge service models, uphold the bottom line of humanistic values, and ultimately position themselves as core nodes of knowledge infrastructure in the AI era, continuing to contribute to the development of human society.

I hope my analysis can inspire your thinking. This is a complex and challenging era, and the future of libraries requires our joint exploration and shaping.

The June 2024 issue of "Library Construction" features a special paper on "The Inheritance and Development of the Spirit of Chinese Public Libraries."