Introduction In the rapidly evolving landscape shaped by artificial intelligence (AI), we are witnessing an unsettling trend: cognitive surrender. In this article, we delve into the impacts of AI on professional skills, significant advancements in stem cell therapy, ongoing challenges in lung cancer treatment, and insights from thought leaders like Cory Doctorow. Join us as we explore these critical themes that underline the intersection between technology and human capability.
Part the First: The Dangers of Cognitive Surrender Recently, we examined an essay on AI that introduced the concept of “cognitive surrender.” This term aptly encapsulates the current predicament regarding our dependence on technology. A recent article in Nature titled Is AI ruining our skills? Early results are in – and they’re not good indicates that reliance on AI tools is diminishing the skills of professionals such as doctors and software engineers:
As more professionals rely on AI tools, could their hard-earned abilities deteriorate?
This concern resonates deeply with medical specialists, computer scientists, and other workers. For instance, surveys reveal that 70% of nurses and 77% of physicians express worry about losing their skills due to an over-reliance on AI systems.
Such fears may be warranted. Preliminary evidence indicates that ‘deskilling’ driven by AI is emerging in fields like medicine and computer science. Researchers are actively engaging in discussions about preserving crucial human expertise in this AI-dominated era.
This issue isn’t new; it dates back to last October and remains unsurprising:
Consider a study involving Polish physicians specializing in endoscopy, where AI tools can analyze colonoscopy images in real time to flag precancerous lesions. These specialists, who each had performed over 2,000 colonoscopies, experienced a notable decline in their detection skills on non-AI assistance days. Before the introduction of the AI tool, they identified adenomas in 28.4% of colonoscopies; afterward, this rate dropped to 22.4% in the absence of AI support.
While a difference of around 6% might not seem alarming, imagine the implications for your gastroenterologist when it comes to identifying potentially cancerous polyps. Not an appealing thought.
Turning to computer science:
At the AI firm Anthropic in San Francisco, researchers conducted a randomized controlled trial with 52 software engineers assigned a basic coding task. Participants could search online for guidance; half also utilized an AI assistant.
Following the task, a quiz gauging their learning revealed that AI-assisted participants scored markedly lower—only 50%—compared to 67% for their non-AI counterparts. Notably, those who had engaged with AI struggled particularly with error diagnosis, indicating a failure to grasp critical coding concepts they had just employed.
This study was shared on the preprint server arXiv prior to peer review. (Please note, referencing arXiv implies the findings are yet to undergo formal review.)
Thus, software experts faltered from a D-minus (67%) to an F-minus (50%) by essentially relying on AI. This trend mirrors the experiences of medical students who depend on online tools that promise effortless studying—a precarious strategy when facing the rigor of their board exams after two years of medical training.
As articulated in the article, we witness a form of “Outsourcing cognition,” which is another way of describing “cognitive surrender.” Education is often misconstrued as merely knowing where to find answers, a notion that remains flawed despite our immediate access to vast information via technology.
The implications? Potential disaster. Accountants may lose fundamental skills, and many will forget how to read a map. Furthermore, your doctor could miss as many as one in five early tumors. The consequences could be dire. One wonders if Sam Altman, in his later years, will prefer a skilled doctor over one reliant on AI for insights during a colonoscopy. One-in-five might work in gambling, but applying it to health is a dangerous gamble.
This trend towards deskilling threatens the future of science, despite any superficial gains in productivity or expansion. Genuine scientific discovery often comes from recognizing anomalies or asking innovative “what if” questions. Max Delbrück famously asserted that experiments must retain room for unexpected outcomes. If scientists lose sight of the data due to excessive reliance on AI, they may become unable to identify the unexpected or even to interpret what should be anticipated. Intuition plays a key role in the labs, libraries, and archives of scientists, literary critics, historians, and novelists alike. This disconnect will not end well.
Part the Second: Remarkable Advances in Stem Cell Therapy An exciting experiment that predates the AI era shows remarkable progress in stem cell therapy. A recent report highlights that stem cells banish severe autoimmune disease for 15 years:
Two individuals with a rare, life-threatening autoimmune condition achieved over 15 years of remission after undergoing stem cell transplants. The research, published in Med, indicates that this experimental approach warrants further clinical trials.
The patients suffered from neuromyelitis optica spectrum disorder (NMOSD), where immune cells produce antibodies that lead to severe spinal cord damage, causing debilitating symptoms. Traditional treatments were ineffectual for them.
Post-transplant, the male patient’s neurological function improved, and he led a normal life, ultimately becoming a father. The female patient regained significant use of her arms and no longer requires medication for symptom management.
This procedure involved chemically obliterating each patient’s immune system and replacing it with compatible cells from a donor—a complex process with notable risks, including infection and potential cancer development. Yet, the alternative remains far more daunting, justifying the pursuit of such clinical experiments. Historically, when stem cell transplants became an option in the 1990s, expectations exceeded reality. However, successful cases like these provide a glimmer of hope.
While progress in stem cell therapy is evident, further understanding remains necessary. The underlying research article, albeit locked behind a paywall, is titled Long-term remission of neuromyelitis optica with allogeneic hematopoietic stem cell transplant. Although only two cases demonstrate success, ongoing advancements in techniques are likely to yield substantial results. Furthermore, this study was conducted in Italy—a country known for its thorough and meticulous research approach. Whether similar outcomes could be replicated at facilities like the NIH remains uncertain as we look to the future.
Part the Third: The Ongoing Battle Against Lung Cancer Clinical oncology has witnessed remarkable progress over the years, but certain cancers, particularly lung cancer, continue to pose significant challenges. As noted in a recent STAT article, lung cancer often stems from well-established causes, leaving a minority of cases classified as “sporadic.” All cancer patients deserve the finest treatments available.
This article highlights that Pfizer’s lung cancer drug failed in clinical trials, a situation I have observed closely due to the drug’s target, which plays a crucial role in cell adhesion and mobility—topics I once researched with my students:
Last Monday, Pfizer announced that an experimental drug aimed at replacing a common chemotherapy for lung cancer did not meet its clinical trial objectives.
High expectations hung on this drug, sigvotatug vedotin, anticipated to succeed docetaxel, a treatment first approved in 1996. Last year, Pfizer’s CEO referred to the drug as a potential growth driver for the company. Despite heightened optimism from analysts, the absence of statistically significant survival improvement over docetaxel means further validation is required.
The target protein, integrin beta-6, is known to be involved in cancer and is upregulated in 90% of non-small cell lung cancer tumors. Despite its targeted approach, the drug failed to achieve meaningful improvements in patient outcomes.
Understanding the science behind this failure illustrates the challenges that remain in cancer research. The roles of complex proteins like integrins in cell behavior are still largely unknown, highlighting the incremental nature of biological science advancements. Identifying promising targets requires a deep understanding of their functions; mere presence in cancer cells is insufficient to warrant effective treatments.
Future research may consider combinations with immune checkpoint inhibitors like Keytruda, though this approach remains uncertain. Furthermore, it is essential to remember Pfizer’s substantial investment in acquiring Seagen—approximately $46 billion—while the realities of research and development continue to unfold.
Part the Fourth: Reflecting on Louis Pasteur’s Legacy Louis Pasteur, often recognized as a chemist rather than a doctor or biologist, made monumental contributions that shaped modern science, including proponents of the germ theory of disease. His innovations came against the backdrop of competing theories such as the terrain theory proposed by Antoine Béchamp. While Béchamp’s ideas were viable before germs were understood, Pasteur’s germ theory ultimately triumphed.
Popular misconceptions aside, Pasteur’s early work focused on the study of crystals—especially those formed during fermentation. His groundbreaking discovery that these crystals were mirror images revealed a profound understanding of molecular structures, laying foundational principles for modern organic chemistry—a contribution later recognized with the first Nobel Prize in Chemistry awarded to Jacobus van’t Hoff.
Further advancements in bioluminescence science illustrate the importance of correct molecular structures. Early synthetic attempts produced an inert product due to incorrect stereoisomer composition—a vital lesson in the significance of molecular structure. Today, therapies leveraging such insights enlighten pharmaceutical developments. Future discussions will explore the implications of “mirror life” created by synthetic biologists, a territory laden with possibility and peril.
Part the Fifth: Insights from Cory Doctorow on AI I had intended to review Cory Doctorow’s new book, The Reverse Centaur’s Guide to Life After AI: How to Think About Artificial Intelligence – Before It’s Too Late, released earlier this week. Yet, Front Porch Republic has already provided an insightful review by Joshua Pauling. Doctorow’s idea of the “reverse centaur” aptly captures the current AI landscape—a horse’s head combined with human legs.
Doctorow argues that the frenzy around AI is less about the technology itself and more about the economic interests that drive venture capital and large tech firms. Companies manipulate narratives to maintain growth; hence, they embed AI features within existing products, often against consumer demand. This artificial expansion ultimately jeopardizes broader economic stability, illustrating that only a handful of firms dominate the narrative in U.S. markets.
The review presents a compelling quote highlighting the absurdity of pushing for driverless trucks as “just a shittier version of a train.” Doctorow emphasizes the need for critics to focus on AI’s financial motivations and call for a more thoughtful evaluation of the technology. As Pauling notes, while Doctorow’s insights may not fully encompass all future scenarios, they provide a solid foundation for understanding the implications of our current tech trajectory.
In conclusion, the interplay between AI and various fields, from medical expertise to groundbreaking therapies, is a double-edged sword. As we navigate through these complexities, we are reminded of the critical necessity for maintaining human skills and understanding. The challenges of cognitive surrender loom large, but they are balanced by remarkable advancements in fields such as stem cell research and oncology. Thank you for taking the time to read this reflection. Until next week, take care and stay vigilant. Time, indeed, marches on.
