From Scarcity to Abundance -AI’s Role in the Future of Education, Healthcare and HR
10-Second Summary
This article explores the transformative role of Generative AI in reshaping education, healthcare, and HR, signalling a new ‘cognitive revolution’.
By democratising specialised skills like personalised tutoring, medical diagnosis, and talent evaluation, AI is fundamentally altering the competitive landscape and economics of these sectors.
The article highlights the ethical, legal, and practical complexities of integrating AI with human expertise, as businesses adapt to a new era of abundance and scalability.
The Cognitive Revolution
This is the fourth instalment in our series on the transformative potential ushered in by generative AI. Previously, we proposed a seven-stage trajectory of AI adoption and integration within enterprises, unpacked the AI Enterprise Tech Stack, and posed five strategic questions that guide enterprise AI engagement and adoption. Our view is that companies integrating this new wave of technology into their core value propositions will be the ones that not only survive but thrive in the years ahead.
In this post, we look at three diverse examples—education, healthcare, and HR—to illustrate a broader pattern that AI is set to catalyse across multiple industries.
Specifically, major technological breakthroughs tend to reshape the dynamics of scarcity and abundance. When technology makes a previously scarce product or service abundant and accessible, the competitive focus shifts. Businesses must then innovate by concentrating on features or services that are still difficult to replicate or automate.
The story of human progress is not just marked by, but fundamentally driven by such transformations. Take, for instance, access to light after sunset: its efficiency has surged a thousand-fold while its cost has plummeted 99.9% since the 1700s(1). Or consider the digital photography revolution: if you’re over 35, you’ve lived through the transition from the scarcity of analog to the abundance of digital photography—a shift that not only changed the industry but our relationship with photographs themselves.
But the industrialisation revolution offers perhaps the most instructive parallel. Before the 18th century, textiles were a luxury, crafted by skilled artisans in a labour-intensive process. As a result, most people owned only a few sets of clothes. The mechanisation of textile production turned this scarcity into abundance, significantly reducing the cost and increasing the availability of clothing. This transformation wasn’t just about making textiles cheaper and more accessible; it also changed the competitive landscape for businesses. With basic clothing now readily available, textile companies had to find new ways to innovate to attract consumers, focusing on aspects like quality, design, and branding. This shift didn’t just transform the textile industry; it also ushered in societal norms around fashion, personal expression, and even hygiene. Over the past two centuries, the industrialisation that began with textiles has extended across most manufacturing sectors, redefining our relationship with nearly all the products and services we consume today.
AI heralds a cognitive revolution in knowledge work, akin to the seismic shift brought about by the industrial revolution in manufacturing. Just as the industrial revolution facilitated a shift from artisanal crafts to mass-produced artefacts, AI has the potential to make specialised knowledge—such as personalised tutoring, medical diagnosis, or talent evaluation—abundant and accessible. In these sectors, businesses will need to innovate new features and services to maintain a competitive edge. In the process, AI not only opens new possibilities but also challenges the existing structures—making traditional doctor-reliant diagnoses, teacher-reliant education, and panel-reliant hiring processes increasingly obsolete. Now let’s take a closer look at these examples.
Khanmigo: An AI Guide at Your Side
Education has long grappled with the challenge of scalability, particularly when it comes to personalised instruction. Research has shown that students benefit immensely from one-on-one interaction. This phenomenon, known as the 'two sigma problem,' highlights the difficulty of scaling such benefits in traditional classroom settings (2).
Khan Academy has been at the forefront of addressing this issue with its asynchronous learning model. However, the recent integration of AI, specifically GPT-4, has allowed them to take a leap forward with KhanMigo. This AI-powered personal tutor provides real-time, interactive guidance, offering a significantly enriched learning experience. It allows Khan Academy to broaden its subject offerings while still maintaining its core value proposition: a free, world-class education for anyone.
KhanMigo isn't just a cute add-on feature; it represents a watershed moment in how AI could transform educational settings. With features like real-time interactive guidance, KhanMigo enables personalised learning on a mass scale, something that has been a theoretical ideal but practically challenging to implement. It goes beyond mere assistance—it has the ability to understand computer code, offer reading support, and even engage in collaborative storytelling. Moreover, KhanMigo comes with a 'teacher mode' to assist educators in crafting lesson plans tailored to individual learning paths. Thus, the AI serves as a collaborator, not a replacement, for human educators. Its ongoing development, including planned features like automated essay support and AI-driven study planning, promises to continue pushing the boundaries of what is possible in edtech.
KhanMigo exemplifies the cognitive revolution AI is bringing to education. By making personalised tutoring—an area once limited by human bandwidth—both abundant and accessible, it's shifting the educational paradigm and forcing a rethinking of how we teach and learn.
Med- Palm 2: Digital Diagnosis on Demand
Healthcare, like education, faces its own distinct set of challenges, but providing early, high-quality diagnoses at scale ranks among the most pressing. Several factors complicate this, including the prohibitive cost of medical care in many regions, limited access to quality healthcare in rural areas, and the inertia that often delays people from scheduling doctor's appointments.
Google's foray into AI-powered healthcare solutions has been years in the making, with innovations ranging from detecting breast cancer (3) to assisting in genome sequencing (4) . However, their most groundbreaking effort to date is Med-PaLM 2. This refined model has been ‘prompt-tuned’ on a curated set of expert medical demonstrations. While AI systems had long struggled to get beyond the 50% pass mark on medical licensing exams, Med-PaLM 2 achieved an 85.4% accuracy, rivalling the diagnostic capabilities of highly competent human doctors. The potential impact on healthcare diagnostics, although raising important ethical and regulatory questions, is enormous.
What sets Med-PaLM 2 apart is its adaptability. It isn’t just a standalone product; it also serves as a foundational layer that can be integrated into niche healthcare applications like mental health, sports medicine, and geriatric care. Imagine it as an ever-present, on-demand digital triage nurse accessible through web or mobile applications. In this role, Med-PaLM 2 could kickstart the diagnostic process, capturing more comprehensive data over time than traditional methods.
Additionally, healthcare providers can employ Med-PaLM 2 as a diagnostic co-pilot, offering a secondary perspective that can either affirm or challenge initial assessments. This 'AI second opinion' can contribute to a more holistic patient care strategy, supplementing human expertise rather than replacing it.
The broader implications of Med-PaLM 2 are transformative. It challenges the traditional economics of healthcare, which have been governed by the scarcity of specialised diagnostic skills. Technologies like Med-PaLM 2 have the potential to make these skills as ubiquitous as internet access. In an echo of Adam Smith's paradox of value, it's a shift that changes where value is created in healthcare. As diagnostic capabilities become more automated, healthcare providers may find their competitive edge in still-scarce services such as personalised treatment plans, specialised expertise in complex conditions, and the quality of patient experience. It's in these nuanced, human-centric areas that healthcare providers may redefine their competitive landscape, ushering in a new era of abundant diagnostic capabilities.
Sapia and Vervoe: Automating the Hiring Process to Make it Faster , Fairer and More Effective
Hiring is hard. Firstly, the process is often inefficient, particularly in today's remote work landscape where the pool of potential candidates can stretch across the globe. Manually sifting through thousands of applications for a single role is both time-consuming and impractical. Secondly, traditional methods like CV reviews and interviews don't always yield reliable insights into a candidate's true capabilities or potential for job success (5). Lastly, the human factor introduces various forms of unconscious bias into the selection process, notably the 'mirror effect,' where we are inclined to favour candidates who resemble ourselves (6).
Much like personalised maths tutoring and medical diagnosis, the intricate task of selecting the right candidate for a role has long been viewed as a domain requiring the nuance and adaptability of human judgement. However, this is another area where AI introduces compelling possibilities for increased efficiency and scalability.
Sapia is one example of how AI is being applied to effectively automate the recruitment process. Sapia uses natural language processing through a chat interview to assess candidates based on a set of standardised questions. Interviews are untimed, so candidates can take as long as they want to answer. The AI analyses candidate responses on a range of attributes spanning communication skills, personality traits, and behavioural competencies, all without revealing demographic characteristics like gender, age, or race. Shortlisted candidates then progress to an automated video interview, which is also untimed and allows multiple attempts.
Vervoe takes a different approach with their AI-powered skill testing platform. Focused on the premise that hiring should be about merit rather than background, Vervoe uses machine learning to create tailored skills assessments that are instantly auto-gradable. This allows companies to test candidates for any skill and automatically grade their responses at scale. The platform is designed to identify candidates who can actually perform the job, bypassing the traditional reliance on CVs and cover letters.
Although Sapia and Vervoe are focused on making the hiring process faster, fairer and more effective, AI will likely transform the entire lifecycle of HR - hiring and onboarding to learning and development, performance management and even offboarding. Yet, as with education and healthcare, the ethical and legal intricacies of blending human and machine intelligence in HR remain challenges that are still being navigated.
AI, like major technological breakthroughs in the past, will alter the dynamics of scarcity and abundance. Here we’ve explored the first inklings of how this might look across education, healthcare and HR, but the future is unpredictable, and the way each ecosystem changes will inevitably contain some surprises.
The key insight is that AI will do more than just make existing processes more efficient; it heralds a cognitive revolution that will fundamentally alter the competitive landscape. This shift requires businesses to discover new, hard-to-automate sources of value as the basis for competition. In this way, the next chapter in human progress is set to unfold.
REFERENCES >
(1) https://ourworldindata.org/light-at-night
(2) Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring. Educational researcher, 13(6), 4-16.
(3) Using AI to improve breast cancer screening (blog.google)
(4) Genomics efforts from Google Health (blog.google)
(5) Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological bulletin, 124(2), 262.
(6) Hiscox, M. J., Oliver, T., Ridgway, M., Arcos-Holzinger, L., Warren, A., & Willis, A. (2017). Going blind to see more clearly: Unconscious bias in Australian Public Service shortlisting processes. Behavioural Economics Team of the Australian Government.