AI's impact on jobs: Why schools must prioritize soft skills
The rapid advancement of generative artificial intelligence is prompting K-12 educators to fundamentally reassess the core competencies students will need for the future. For decades, the most lucrative careers increasingly revolved around intellectual tasks, particularly in science and technology. However, with the widespread adoption of generative AI, this paradigm is shifting. Employers are now signaling their intent to automate certain professional roles, raising questions about the future demand for creative and analytical workers like computer programmers, and the viability of many entry-level positions within the knowledge economy.
This profound change affects not only the workforce but also K-12 teachers, who have long focused on preparing students for white-collar professions. Families, too, are grappling with anxieties about the skills their children will require in an economy permeated by generative AI. As an education policy professor who has researched AI’s impact on employment, and a former K-12 teacher, I believe the solution for both educators and parents lies in understanding what AI cannot — and likely will not — be able to do.
Previous waves of automation primarily displaced routine and manual jobs, thereby amplifying the earning potential of intellectually demanding work. Generative AI, however, operates differently. It excels at pattern recognition, allowing it to simulate human coding, writing, drawing, and data analysis. This capability renders the foundational levels of these occupations susceptible to automation. Conversely, because its output is based on patterns in existing data, generative AI struggles with intricate reasoning challenges, and even more so with complex, open-ended problems whose solutions depend on numerous unknowns. Furthermore, it possesses no innate understanding of human thought or emotion.
This critical distinction suggests that “soft skills”—the attributes enabling effective human interaction and self-awareness—are poised to become paramount. These skills are integral to solving complex problems and collaborating with others. While qualities like conscientiousness and agreeableness are often considered personality traits, research indicates they are, in fact, developable emotional competencies that can be taught.
The encouraging news is that these crucial soft skills can be integrated seamlessly into the teaching of traditional subjects like mathematics and reading, areas for which teachers are already held accountable, using familiar pedagogical techniques. For instance, teachers often use “exit tickets” at the end of a lesson—brief written reflections or questions about newly learned concepts. These can be adapted to help students hone their emotional and social competencies alongside their academic learning. Teachers might offer prompts that encourage reflection on moments of intellectual resilience, emotional control, or interpersonal understanding, such as: “Describe a time today when you helped someone,” or “Tell me about someone who was kind to you today, and how they showed kindness.” Another effective prompt could be: “Recall a time this week when you learned something that felt incredibly difficult. How did you overcome that challenge?”
The primary aim of such exercises extends beyond merely boosting student mood or engagement, though these are valuable byproducts. The core objective is to help students recognize that their emotional responses to external circumstances are within their sphere of influence. Enhanced awareness of their own emotions has been shown to predict children’s ability to manage frustration, anticipate and perceive the emotions of others, and collaborate smoothly with peers. All of these are vital workplace skills that will undoubtedly grow in value with the proliferation of generative AI.
Teachers can also engage students in solving “messy problems” for which answers are not immediately known. For example, as elementary students learn to calculate perimeters or volumes, they can work in groups to measure large or unusually shaped objects around the school. The emphasis should not just be on the correctness of their answers, but on how they framed and approached each problem. This real-world application of knowledge, often termed authentic assessment, can be implemented across any discipline. Examples include analyzing soil slopes and moisture levels on school grounds to propose landscaping solutions, creating and pilot-testing video campaigns for social causes, or reimagining historical events by considering how different leadership choices might have altered outcomes and their modern policy implications. Teaching children to unravel complexity helps them distinguish between seeking textbook answers and exploring possibilities when the optimal solution is uncertain. Solving novel, complex problems will continue to pose significant challenges for AI, not only due to the numerous steps and unknowns involved but also because AI lacks our intuitive grasp of context and human emotion. Even in the long term, countless variables that humans instinctively understand will remain difficult for algorithms to replicate.
The most frequent concern I hear from teachers regarding technology is students using generative AI to complete their assignments. This behavior stems not from malice but from human nature; we are driven by efficiency and reward, often taking shortcuts on tasks that seem dull or daunting to prioritize more gratifying activities. However, when students are in the crucial phase of developing new skills, delegating work to AI is a significant misstep. By making slow processes fast, AI can inadvertently undermine learning, as true skill acquisition often requires sustained effort through challenging tasks.
Therefore, I believe educators must safeguard the classroom as a space where foundational skills are acquired gradually, in collaboration with peers. For many lessons, this might entail revisiting traditional learning methods from the pre-computer era, where students wrote assignments by hand or presented their work orally, learning to anticipate and respond to diverse viewpoints. If students are permitted to use digital automation tools, they should be prompted to reflect deeply on how they utilized them, what insights they gained, and, crucially, which skills they were unable to practice—such as spelling, long division, or bibliography formatting—by delegating the work to the tool.
While no one can definitively predict the precise future of work in an AI-enabled economy, and experts disagree on which skills AI will augment versus replace, the foundational competencies of mathematics and reading will undoubtedly remain important. Crucially, the innately human skills of self-awareness and social interaction will become even more vital. Perhaps the most critical skill schools can impart to children today is the self-awareness to prioritize genuine learning over shortcuts, and to refrain from delegating tasks to machines until they are capable of performing those tasks themselves. Furthermore, the ability to collaborate effectively with others to dissect and solve difficult problems will become increasingly indispensable. An AI-enabled society will not be one devoid of complex challenges. Even as the labor market reconfigures itself, I am confident that opportunities will abound for those who can work effectively with others to tackle the significant challenges that lie ahead.