Dear Friends and Colleagues,
Every transformative technology arrives with a prophecy. The railroad would collapse geography. The printing press would shatter the monopoly of the learned. The internet would democratize knowledge. Each of those technologies did reshape the world — just not in the ways, nor at the speed, that their champions promised. The AI prophets are no different — and it’s prudent to remember many of them have a vested interest in hyperbole.
A voice worth heeding about AI and the future of work belongs to David Autor — MIT economist, co-director of the NBER’s Labor Studies Program, and the scholar The Economist called “the academic voice of the American worker.” Autor and co-author Neil Thompson frame AI’s promise as the extension of human expertise: a force multiplier for judgment, creativity and knowledge. Its peril, they argue, is not mass unemployment but something more corrosive: the devaluation of expertise itself. Put differently, AI commodifies precisely what conventional schooling reliably produces — coverage, recall and the completion of well-defined tasks.
The labor market is registering this shift. The World Economic Forum’s Future of Jobs Report 2025 found that analytical thinking is the skill employers value most, followed by resilience, adaptability, leadership and creative thinking. What the economy wants is not people who execute routine tasks faster than machines — that race is over.
The message for educators is unambiguous. Developing the skills machines can’t replace must become a main product, not a byproduct, of schooling.
To be clear, this is not an argument against developing academic knowledge. Reading, writing, quantitative reasoning and disciplinary content are not enemies of collaboration, communication and critical thinking. They are the substrate on which those capacities grow.
The good news is that schools across the country are already solving for this.
- At Northern Cass School District #97, 20 miles north of Fargo, North Dakota, a district of 700 students has spent about a decade embedding durable skills into every course — content mastery and rigorous inquiry live side-by-side in the same lesson, the same assessment. Neither is optional.
- At Liberty Public Schools in Kansas City, a district of 16,000 has built an unusual credentialing architecture — portfolios, digital badges and curated evidence of student work — that travels with young people into college applications and job interviews, preparing them for the new economy.
- At Cañon City Schools in Colorado, no senior graduates without completing an 18-month capstone project, presenting their work to a panel of outside experts and their entire graduating class.
These school systems share two common insights. First, naming a skill is not the same as developing it. A portrait of a graduate that hangs in the hallway but doesn’t live in the rubric, the transcript or the graduation requirement is a decoration, not a strategy. Second, developing skills at scale requires shared, science-based definitions and clarity about how those skills grow over time. Collaboration is not one thing — it encompasses perspective-taking, conflict navigation, role clarity and shared accountability. These subskills have developmental trajectories. They can be sequenced, assessed and taught.
What these schools demonstrate in practice, Carnegie is working to make possible at broader scale — building skills standards in partnership with educators and employers across the country and translating those standards into tools for curriculum designers, assessment developers and teacher preparation programs — so that when we say “communication,” “collaboration” or “critical thinking,” we mean the same things across the ecosystem. In early 2027, we will publish science-based frameworks for the nine skills that best predict success in school, work and life.
Clearly, AI can be built to amplify human expertise — but only where genuine expertise exists to be amplified. For young people today, that condition will be met if we take both disciplinary knowledge and durable skills seriously: developing students with the judgment to know when a machine is wrong, the creativity to ask questions not yet asked, the adaptability to change course when a strategy fails, and the relational intelligence to build trust, hold complexity and bridge divides.
Schools can build those capacities — some already are.
In partnership,
%20(2).png?upscale=true&width=450&upscale=true&name=CAR%20logo_INTER_Black@4x%20(1)%20(2).png)
.png?upscale=true&width=600&upscale=true&name=CFN_600x400_Feb24_v4%20(2).png)
