If we consider the vast material in the U.S. Library of Congress as "intelligent," then it becomes reasonable to classify augmented large language models (LLMs) like ChatGPT, GROK, and DeepSeek as intelligent as well.
These systems have been augmented with such additions as web search and mathematical abilities. Augmented LLMs distill massive amounts of information into useful responses, making knowledge more accessible than ever before.
But if intelligence includes creativity or understanding, then neither the Library of Congress nor augmented LLMs qualify. They do not understand, and they cannot generate original insights. Creativity and understanding remain exclusively to human exceptionalism.
Newton and the Shoulders of Giants
Isaac Newton famously said, “If I have seen further, it is by standing on the shoulders of giants.”
He was likely referring to predecessors like Galileo, Kepler, and Descartes. Newton built on their work, but his development of calculus and classical physics were unprecedented contributions above and beyond — products of human creativity.
Today, augmented LLMs make standing on intellectual shoulders easier. Newton's personal library contained about 2,000 volumes. By contrast, the U.S. Library of Congress now holds over 178 million items, and the U.S. National Archives and Records Administration (NARA) contains billions of documents, photos, maps, and recordings. These archives form an enormous foundation of knowledge.
Augmented LLMs can process text akin to the Library of Congress and NARA. They are unparalleled tools for searching, summarizing, and retrieving data.
But they do not understand the meaning of what they process. Nor can they create. The intellectual giants of the past provide the foundation — but it takes the human intellect to build something new atop it.
Can AI Ever Be Creative?
This question is addressed by Selmer Bringsjord’s Lovelace Test, which states that a program is creative only if it produces an output beyond the intent or explanation of the programmers. No AI system has passed this test.
Even the transformer model introduced in the 2017 paper “Attention is All You Need” which underlies LLMs, behaves entirely within the expectations of its designers. While powerful, it does not generate truly new ideas — it merely recombines known ones. AI can only "think" inside the box it was trained in.
AI as a Mathematical Tool
Claims that AI is creative often rest on fuzzy examples. A true, inarguable benchmark of creativity is solving longstanding open problems in mathematics — like the Riemann Hypothesis, for which the Clay Mathematics Institute offers a million-dollar prize. (The Riemann Hypothesis was referenced in the Academy Award-winning 2001 film A Beautiful Mind.)
Mathematicians are already using machine learning to visualize complex patterns and develop proof assistants to check the validity of their arguments. AI can brute-force search within a finite solution space, but the creative act lies in narrowing that space in the first place — a uniquely human achievement.
Such AI tools assist. They do not originate creatively.
Even though forecasts about AI being creative in mathematics have been made, no AI has independently solved any major open problem in mathematics. Every recent breakthrough in math has involved human creativity. Consider, for example, the list of recent mathematical breakthroughs in Quanta Magazine.
Until an AI passes the Lovelace Test, claims of machine creativity cannot be made.
The Elusive Goal of Artificial General Intelligence
The concept of Artificial General Intelligence (AGI) is fuzzily defined. If AGI merely means easier access to knowledge — climbing Newton’s “mountain” faster and with greater ease — then AGI is possible.
Some argue it is already here. But if AGI means replicating human creativity and understanding, no.
But what about computers of the future? The Church-Turing Thesis tells us that all computations by modern machines are, in principle, equivalent to those of a Turing machine from the 1930s.
Faster hardware does not expand the boundaries of what machines can fundamentally do. This is also true of quantum computing.
LLMs are remarkable tools for navigating vast knowledge bases. They retrieve. They organize. But they do not create nor do they understand.
The Final Distinction: Tools vs. Minds
Ultimately, LLMs are tools — powerful, efficient, and capable of augmenting human accomplishment in extraordinary ways. But to call them intelligent in the same way we describe human minds is a mistake.
True intelligence lies not in the volume of stored knowledge, but in the uniquely human capacity to transform knowledge into something new. It lives above the shoulders of giants where the creative mind takes flight.
Robert J. Marks Ph.D. is Distinguished Professor at Baylor University and Senior Fellow and Director of the Bradley Center for Natural & Artificial Intelligence. He is author of "Non-Computable You: What You Do That Artificial Intelligence Never Will Never Do," and "Neural Smithing." Marks is former Editor-in-Chief of the IEEE Transactions on Neural Networks. Read more Dr. Marks' reports — Here.
© 2025 Newsmax. All rights reserved.