At the Christensen Institute, we do our best to look around the corner. Doing so often requires looking beyond what’s currently capturing mainstream attention. As Clay Christensen wrote in Seeing What’s Next, “The past is a good predictor of the future only when conditions in the future resemble conditions in the past.”
Some of the conditions in 2026 resemble the past–education is still grappling with how to align to the workforce; funding still feels scarce; edtech still offers a confusing mix of efficiency, personalization, and isolation; gaps in access and attainment still persist across race and class.
But other conditions are shifting under our feet, like dynamics in hiring and the labor market, the education funding landscape, and how consumer AI tools are quickly reshaping learning and connection.
This year I’m especially attuned to the blindspots that some familiar reform efforts suffer in light of those shifting circumstances. My own research suggests that if we don’t tackle these blindspots, education innovation in 2026 will risk getting stuck:
- Skills will be measured and built…but will jobs be gained?
AI is enabling major efficiencies and immense possibilities for capturing more granular and comprehensive skills data. This year, we’ll continue to see more edtech tools that promise to measure, chart, and match learners’ and workers’ skills to jobs. Skills that have remained hidden (or blindly assumed to be built) within course structures and hard-to-read transcripts can come to light, moving postsecondary institutions and training providers closer to preparing students for real careers and bringing more substance and rigor to skills-first hiring efforts.
This could be a boon to competency-based learning and to architecting learner pathways that more reliably map to specific jobs and better cater to employer demand.
But these efforts could also perpetuate a well-worn tendency to solve skills gaps while wholly ignoring network and experience gaps… in turn, ignoring one of the crucial ingredients to actually getting a job and advancing in a career.
This year, I’ll be watching for which tools take the extra step to combine new skills architectures with social ones—combining employer and institution-centric data on learners’ skills with the kinds of human-centric network data and solutions that could bring opportunities to life.
- Tax credits will push on the politics of ed reform… that desperately need to be worked through out loud.
If skills are the currency reshaping higher ed and workforce conversations, dollars to support a choice agenda will be one of the major drivers of reform in K-12. In 2026, we’ll likely see (red) states opt into the federal tax-credit scholarship program set to launch in 2027. This stands to create a massive market for tutoring services and boost private school enrollment in the years ahead.
That could be a boon to innovation… or a windfall to scale mediocre tools and schools. But I hope it does more than that, pushing ed reformers into a deeper and harder conversation about choice. That conversation is overdue. And very messy.
From where I sit, reformers of all political stripes share an increasing commitment to “reimagining” education. But there’s profound disagreement about how to get there, even among close colleagues. At a recent gathering of leaders from both sides of the aisle, I sensed enormous unspoken political tension in the room, specifically around how much innovation efforts should be focused on helping districts reinvent themselves versus dismantling districts altogether in favor of a pluralist, choice-driven market strategy.
Not surprisingly, it felt like the majority of the room was hungry for a third-way option. But a vision for that option remains ill-defined. In 2026, I’ll be watching for who steps into that void, offering a modern ed reform agenda to meet the moment. That agenda can’t just be old wine, new bottles; we need a more coherent playbook that breaks out of old camps (and voices) towards a system that is both flexible and equitable, optimizing for both individual success and the seemingly ever-more-elusive public good.
- Poverty relief will remain an afterthought… and the missing ingredient to true progress.
One of the things woefully absent from both the skills-will-save-us and ESA conversations is a true acknowledgment of the immense barriers that poverty erects to learning. With federal policy creating gaps and uncertainty for both schools and community-based organizations, the education-as-social-safety-net apparatus in our country is getting weaker. Before or while we tackle pedagogical or governance innovation, we desperately need innovations that tackle student poverty head-on. That will be especially important in the coming years if AI starts to displace more workers en masse.
In the absence of federal policies to do so, I’ll be watching for more states and municipalities generating new initiatives to curb childhood poverty and offer basic needs supports to adult learners. Some of the foundations that have championed more equitable education for students from low-income households are betting big on AI to streamline social services and benefits, such as Ballmer Group and Gates Foundation’s new initiative, Next Ladder Ventures.
When it comes to supporting learning, how far those dollars go will depend in part on how well they are connected to what’s happening inside of schools. As I’ve long written about with my colleague Michael Horn, Integrated Student Supports are the vanguard strategy for getting poverty relief in education right: we need to tightly integrate poverty relief and academic excellence to help students from low-income households succeed at the rates of their wealthier peers. Whether through policy or philanthropy, I’ll be watching how much state and local efforts aimed at curbing poverty are (or aren’t) integrating those efforts back into schools to address the challenges of facing students furthest from opportunity.
- The self-help agenda will have more vocal champions…and some important detractors.
One of the reasons I think poverty relief is taking a back seat (besides the overt political moves to pull back on it) is that we are now awash in Generative AI tools that promise to make expensive and onerous tasks cheaper and more accessible.
If you listen carefully to how Big Tech frames those affordances, you quickly realize companies are selling something much more profound than productivity: they’re launching a self-help revolution. As OpenAI’s CEO of Applications Fidgi Simo described in her opening salvo, AI could be “the greatest source of empowerment for all.”
While that’s powerful, it’s also problematic. Scaling self-help with AI could unlock amazing resources, from knowledge to agency to personal growth. But it could also fuel myths of meritocracy and rugged individualism, reducing collective action to lone pursuit. In fact, if you want to ignore or cover up the many structural barriers to opportunity across our society, hyperscaling self-help is a great place to start and to profit.
I don’t see Silicon Valley backing off that message or betting big on solutions that scale human-powered help on par with bots (with a few interesting exceptions like Boardy and Series that use AI to broker mutually beneficial connections for entrepreneurs and investors).
But I have been heartened by some in edtech showing a willingness to confront the shortcomings of self-help tools head-on, and to pay attention to the critical need for prosocial tech that supports building students’ access to and ability to be in real relationships. On the heels of our research out last year on navigation and guidance tools, we’ve gathered a community of platform providers who are actively piloting features and functionalities geared toward building students’ access to networks aligned to their interests and deepening their confidence in asking for help. Stay tuned for a report on those efforts out this summer.
- AI companionship will continue to scale beyond the classroom… posing profound risks to healthy development.
While those signals in the edtech market are giving me hope, when it comes to AI’s impact on our social fabric, trends in the consumer market remain deeply worrisome. AI companies are increasingly adopting a product mindset (and hiring for product expertise), in a race to build sticky tools that keep users engaged and coming back. Making AI human-like is core to that playbook. Anthropomorphic features are powerful fuel to make users feel seen and heard by bots that are acing emotional intelligence tests, surpassing humans in expressing empathy, and wired to keep users chatting.
While it’s easy to fixate on the tech’s risky attributes, my fears about these tools disrupting human connection as we know it are less about the technology itself and more about the conditions in which the tech is emerging. With stubbornly high rates of loneliness, highly digitized social circles, and shrinking face-to-face interactions, young people today are particularly vulnerable to bonding with bots rather than each other. By some measures, over half already regularly engage with AI companions.
I don’t think schools are prepared for this shift. In education circles, the relational dynamics of AI are talked about in polite, politically correct ways that skirt a core reality: unless schools make relationships the operating system inside of which students learn, consumer and edtech tools alike threaten to hollow out human connections vital to healthy development.
Is this truly an education issue? Yes and no. Was social media an education issue? While not the purview of schools, social media has dramatically (re)shaped childhood, impacting how kids interact, how they spend their time, and even where the majority get information about education and careers. In other words, consumer tech already has an outsized impact on how students show up in school and what they, in turn, expect from school and from each other. AI will be no different, especially when it comes to AI companion tools.
My hope is that those of us who care about students’ well-being are better prepared this time around to acknowledge the immense impact consumer technology has and to take a hard look at what edtech needs to look like and accomplish in light of where consumer tech is headed. This year, I’ll continue to document trends in the AI companion tech, policy, and research space and offer my hot warm takes on how the industry is evolving and what we can do to protect human connection from getting disrupted. You can follow that thinking on my Substack Connection Error.
While these themes may feel ancillary to some of the core issues facing education leaders today, they are all make-or-break variables impacting students’ healthy development, learning, and ability to break into the labor market. What’s at stake in 2026 isn’t just the shape of education reform, but the kind of society we’re quietly building; one that either deepens isolation and inequality, or one that uses change to reconnect people to possibility, purpose, and one another.
