quarta-feira, 8 de maio de 2024

The Inquiry Paradox: AI’s Divergent Reasoning and the Human Quest for Meaningful Questions



Reflecting on a profound statement uttered by Professor Sobrinho Simões (named the most influential pathologist globally by the British magazine "The Pathologist,") during an event at the Belém Palace, the official residence of the President of Portugal, in 2017, wherein he posited that "the true hallmark of the human species lies not merely in our ability to provide answers, but rather in our capacity to pose meaningful questions," I find myself compelled to raise the following inquiry: Why is it that when tasked with identifying the most paramount question, two cutting-edge artificial intelligence systems, ChatGPT and Microsoft Copilot, offer disparate responses, as illustrated below?

ChatGPT's Response: That's quite the existential question! The most relevant question may vary depending on context and individual perspectives. However, a fundamental question that resonates across cultures and disciplines could be: "What is the meaning of life?" It's a timeless inquiry that has puzzled philosophers, theologians, scientists, and thinkers throughout history. The pursuit of understanding our purpose and existence seems to be a central theme in human thought and exploration.

Microsoft Copilot's response:  That’s a profound question. If I were to select one question that could be considered universally relevant, it might be: “What can we do to improve the well-being of all?” This question encompasses a range of important aspects such as health, education, equality, and the environment, reflecting the interconnectedness of our global community. It encourages us to think collectively and act in ways that benefit not just ourselves, but everyone.

Upon posing the identical question to the Scopus AI model (that I previously discussed in this blog), the resultant answer proves somewhat unremarkable. However, nestled within the Foundational Documents section, emerges a publication titled "Attention is all you need," boasting an impressive accolade. In merely seven years, this publication has garnered over 50,000 citations in the Scopus database. Authored by eight former Google employees, it unveils an exceptionally groundbreaking discovery, pivotal enough to catalyze their departure from the company, subsequently founding ventures collectively valued at $4 billion https://www.japantimes.co.jp/opinion/2023/07/19/commentary/world-commentary/google-ai-discovery/

P.S. - Reflecting on the 20th century, prior to the emergence of Google, the notion that corporations could rival the world's top universities in producing essential knowledge would have been met with skepticism. However, this once far-fetched idea has solidified into an undeniable reality. An article featured on page 56 of The Economist magazine's February 4 to 10, 2023 edition underscored this paradigm shift, revealing that Amazon and Meta “respectively produce two-thirds and four-fifths as much AI research as Stanford University". Furthermore, the scholarly output of Alphabet and Microsoft exceeded that of Stanford. In fact, it was within this latter company that one of the most cited articles of all time was authored: 'Deep residual learning for image recognition.' This seminal work has garnered over 140,000 citations in the Scopus database to date.   In the existing paradigm, universities are facing a potentially dismal future, unless they can pinpoint and harness their innate abilities that distinguish them from what commercial firms are unwilling, unable, or inherently incapable of undertaking.