AI: Vulnerabilities Induced by Change

We are unprepared for the storms ahead and unwilling to redesign the vessel. - Philosopher Laureate David Van Reybrouck

Artificial intelligence (AI) is likely to threaten tasks that are routine and mechanical: scripted, repetitive, follow a predictable pattern and where the inputs and outputs are purely digital text in and text out. Let us look at some of the jobs that are expected to face disruption.

Software Engineers

Companies aren’t just automating repetitive coding tasks, but are also restructuring teams to focus on AI products. As a result, in the first phase remote workers will feel more of the heat. Next, entry level positions will slowly evaporate because simple coding and debugging tasks - once the proving ground for junior software engineers - are rapidly becoming automated by AI. This is despite some major shortcomings with the tech at present. A recent report in a Times of India report states that in 2024, even students from prestigious Indian engineering colleges struggled to find employment. The report claimed that only 10% of 1.5 million engineering graduates are expected to get jobs.

Indian tech firms currently have excess staff in two categories: senior managers and managers. The ‘bulging middle’ typically refers to managers who are between senior leadership and front-line employees. When automation fully sets in, tech firms may eliminate many people in these two senior layers.

Medicine

Repetitive, data-heavy tasks like diagnosing common illnesses, analyzing scans and flagging abnormalities (an accurately) is likely to put pressure on the jobs of front-line doctors (e.g. at sub-centers) and para-medical staff.

For example, Shanghai-based company Synyi AI had recently launched an AI-run clinic in Saudi Arabia called AI Doctor, which independently conducts consultations, diagnoses and suggests treatment via a tablet. A human doctor reviews and then approves each plan. They handle up to 10,000 patients a week, compared with around 100 by a typical human doctor.

Another platform, Microsoft’s AI Diagnostic Orchestrator, was unveiled to diagnose medical cases with 85.5% accuracy, far more than the 20% accuracy of 21 experienced doctors from the US and UK. Unlike doctors who rely on experience and pattern recognition, this tool evaluates multiple variables simultaneously (i.e., lab results, symptoms, imaging) to predict outcomes.

Education

ChatGPT is now ready to help prepare for engineering and medical entrance tests using questionnaires and study materials of competitive tests like IIT. The new “study mode” helps students work through problems step-by-step instead of just getting an answer. This will give serious competition to coaching centers.  

The personalization of education at scale has huge implications for conventional schooling. Personalization has the potential to deliver education to every child depending on her needs and level of learning.

Learning, thus, acquires a new meaning: you don’t move onto the next level until you know the material from the previous one. It’s the opposite of how schools work right now. In traditional learning, time is fixed (everyone gets exactly one school year for arithmetic) while learning is variable (some kids get A’s, others get C’s). In AI-led personalized learning, it’s flipped - learning is fixed (everyone masters the material) while time is variable (some finish in three weeks, others in three months). These AI-powered, self-paced learning plans where students are able to spend master more material in less time are facilitated by fewer teachers.

So, AI is all set to “revolutionize work”. Something similar happened when manufacturing jobs vanished across America’s heartland during the 1970s and 80s. Sociologist Julius Wilson found that the loss of industrial jobs led to not only loss of income, but also social and political upheaval.

In India as well, a significant portion of the workforce is currently engaged in informal, labor-intensive tasks (e.g. packaging workers, delivery staff, textile factory workers) and administrative and support roles (e.g. data entry, scheduling, and customer service). Therefore, it is imperative that we get ready for the cataclysmic changes induced by AI.