Artificial intelligence: The great divide

Abstract

Albert Einstein once said that the true sign of intelligence was not knowledge but imagination. He further emphasized that imagination is the beginning of creation. You imagine what you desire, you will what you imagine, and at last you create what you will. The genesis of anatomy and functionality of the human brain, with its immense capacity for rational thinking and analysis as well as skill development, has remained an enigma which has, however, not diminished its reality, and which has remained open to continuous explorations through the ages. The phenomenal work of Galen, Vesalius, Descartes, Golgi, and Cajal over the past two millennia has provided the anatomical roadmap of the human brain as well as provided evidence for the complex networks involved in the process of human functioning as controlled by the brain.[1] Complex brain imaging, as is possible today, has aided this exploration and provided an objective basis for brain function and the associated clinical manifestations of brain disease.[2] The past decade has been exemplary in the generation of what has now been called the “big data” revolution.[3] This has been the culmination of collecting huge, diverse data sets with the help of highly sophisticated technological tools that have taken the field from the academic laboratories of individual scientists into the realms of the global industry, analogous to migration of the then nascent pharmaceutical industry from the apothecaries and individual physicians to major industrial operations, about half a century ago. Big data comprise databases that cannot be analyzed by conventional statistical modalities. This could be genomic data, generated by the next-generation sequencing (NGS) tools [after the Human Genome Project (HGP) provided the reference human genome over a period of 13 years from 1990 to 2003], it could be imaging data collected from global imaging databases because of the exponential growth in imaging technology in the past 2 decades, or it could be laboratory/anthropometric data. It is important to emphasize that all of it requires clinical correlations for providing appropriate interpretations.[4] Thus, “big data to small information” forms the challenge of this decade, where the associations established by statistical interpretations of big data which took its backbone from machine learning come back again to the physician for clinical correlates and analysis, establishing specific and sensitive “personalized” biomarkers of “normalcy” and “disease,” the hallmark of “personalized” or “precision” medicine. So, the world has come full circle again, from the individual to the community and then back again to the individual, the tangible aim having shifted to mapping each individual human, so that highly specific disease management can be provided, based on “normal” and “abnormal” biomarkers present in each, the overall intangible aim being a quest for “wellness,” “longevity,” and/or “immortality.” But there is a huge cost to this! Notwithstanding the necessity and/or the expense for the actualization of this aim, which involves “creation” of networks simulating a more “predictable” and “controllable” human brain by artificial modalities, now available with humans as a technological tool, which has had a very successful run in translational research, globally, in the past half decade, its success itself has opened up many questions, unasked till now, because no one gave it that much of thought till the concept of “artificial intelligence (AI)” became a tangible reality, affecting lives and livelihoods. The great divide on AI is based on both physical and metaphysical contentions. The metaphysical contentions may be abstract and philosophical, but they are of relevance in the universal scheme of things. Would an individual like to be so completely mapped, such that there are no privacy and freedom anymore? Would it be humans who would take charge of this grand repository of other humans? Who would choose these humans? Further, who would look after the repositories of these humans who look after the repositories of other humans? Would it be the same “created” AI? Would this AI eventually become so efficient that it would replace their creators themselves? What if there was a data leak? Who would be responsible? The moral, ethical, and legal repercussions of these may sound like science fiction today, but the exponential expansion of AI can make this a reality tomorrow.[5] After all, just as the physical world of paper cards, postal letters, landline telephones, telegrams, slide projectors, and camera reels simply disappeared one day, without any noise, what is there to stop the current world from disappearing tomorrow into another reality? But much, much more important than metaphysical contentions, the paradox of AI lies in the fragile and vulnerable physical reality of the world today that we share with other humans, living beings, plants, and the earth per se. The extremely energy-intensive process of training and running AI models, especially the more advanced and “super-efficient” ones that aim toward changing the world tomorrow, has led to significant freshwater consumption, carbon emissions, rare earth extraction, and e-waste dissemination, all of which continue to contribute to the dramatic climatic and environmental existential crisis that humanity is already facing today.[6] While the AI industry has realized this and is trying to find a middle path toward sustainability by decreasing carbon footprints and freshwater consumption, it is estimated that even with these energy-efficient models (which may increase the cost of AI), a freshwater crisis may loom large by 2028–2030, as early as the end of this decade.[7] At this juncture, it seems plausible to think that energy resources could be “borrowed” from the sun, which, as per existing human knowledge, has a life span of another 4–4.5 billion years in its present form. It is interesting to note that there are many articles currently on how AI can be used to improve harnessing of solar energy but none on how solar energy can be utilized as a green footprint to sustain AI for the future generation. This could be a very relevant thought to cultivate in the near future. Mid-decade, all of us humans, be they scientists, technologists, physicians, environmentalists, and colleagues in other professions, in academia or in industry, are at absolute cross-roads. How much are we borrowing from our children today for the envisioned “precision” of tomorrow? Will humans find a sustainable and optimum middle path in time, possibly through the sun? Here is a true question where philosophers and scientists must collaborate and synchronize their thoughts, like the productive temporal collaboration of artist-scientists of the 15th century, who provided the basis of human anatomy before the physicians, anatomists, and surgeons took over, to answer this yet invisible but looming emergency question before the Frankensteinian monster goes out of control and destroys itself and its creators in one go! How much is enough and how fast is the question, and maybe the answer lies in nature itself……. “Nature does not hurry, yet everything is accomplished” - Lao Tsu About the authorDr (Prof) Zia Chaudhuri, MS, DNB, MNAMS, FRCS(Glasg), FICO, PhD(Genetics), FAMS, Gold Fellow ARVO Dr (Prof) Zia Chaudhuri is a Director-Professor of Ophthalmology at ABVIMS and Dr RML Hospital, New Delhi, India with sub-speciality training and work experience in strabismus, pediatric ophthalmology, neuro-ophthalmology, cranio-facial anomalies, nystagmus, ophthalmic imaging and ophthalmic genetics. Dr Chaudhuri is a Fellow of the Royal College of Surgeons of Glasgow, UK (2001), National Academy of Medical Sciences, India (2019), and the Association of Research in Vision and Ophthalmology, USA (Gold Fellow - 2025). She had been a BOYSCAST International Research Fellow of the Department of Science and Technology (DST), Government of India (GOI) at the Jules Stein Eye Institute (JSEI), University of California Los Angeles (UCLA), USA (2011-2012), where she trained in acquisition and analysis of high resolution orbital MRI in strabismus in a unique technology transfer module as well as performed her doctoral work (PhD) on the genetic analysis of strabismus from the Department of Genetics, University of Delhi South Campus, India (2013-2018). Dr Chaudhuri has been principal investigator (PI) of 05 governmentally funded scientific projects on high-resolution orbital imaging in strabismus and myopia, with ongoing analysis based on artificial intelligence tools, as well as, strabismus genetics in the Indian population. She has more than 23 years of experience in teaching undergraduate and postgraduate students in ophthalmology, and over 150 publications, including editing 08 books in ophthalmology. She received the Best International Research Fellow award 2012 from JSEI, UCLA, Best of Show poster award at AAPOS 2012 and the Prem Prakash Award at AIOS in 2006 for her work in strabismus. She is the current honorary editor-in-chief of the journal, Strabismus and is on the editorial board of BMC Ophthalmology for Pediatric Ophthalmology and Strabismus.

Ähnliche Arbeiten