[Title to be determined]

AI, education, and the possibility economy

Schools have spent generations asking students to answer questions formulated by other people. Curriculum authorities select the subject. Teachers define the assignment. Assessment systems establish the expected form of the response. Students may exercise judgment within these boundaries, but they have limited influence over the questions, assumptions, and purposes that structure most of their education.

Generative artificial intelligence has made the limitations of this arrangement harder to ignore. Large language models can explain concepts, summarize texts, solve routine problems, write essays, generate presentations, and produce competent responses within seconds. These systems compete with students on the work that schools assign because much of that work consists of turning information into an acceptable answer.

Schools have responded with restrictions, detection systems, redesigned assessments, and warnings about academic integrity. At the same time, educational institutions are adopting AI to prepare lessons, create exercises, personalize instruction, evaluate writing, analyze performance, and reduce administrative work. The OECD’s 2024 Teaching and Learning International Survey found that 36% of teachers across participating OECD education systems had used AI in their work. In the United States, the figure was 43% (OECD, 2025).

Schools prohibit students from automating their answers while paying companies to automate the school.

The issue extends beyond institutional inconsistency. Student work and institutional decision-making carry different responsibilities, and their uses of AI should not be treated as equivalent. The broader concern is that both applications can preserve an educational structure based on assigned questions, predetermined outcomes, and centralized judgment. Students use AI to complete the intellectual work that schools require. Schools use AI to produce, supervise, and evaluate more of that work.

This response places AI inside an old educational model rather than examining what the technology changes. Competent answers have become faster and cheaper to produce. This development should alter the questions schools ask about intelligence, knowledge, and human contribution. Instead, much of the current discussion concentrates on protecting assignments that machines can now complete.

A different educational orientation begins with a question that generative AI cannot settle on the learner’s behalf:

What do you want to bring into the world?

This question asks learners to use knowledge to create something that does not yet exist. The result might be a scientific explanation, public service, business, teaching method, work of art, community organization, policy proposal, technical system, or new form of coöperation. The learner must identify a purpose, understand the surrounding conditions, gather evidence, develop an approach, test it, and accept responsibility for the consequences.

This orientation provides the basis for what I call the possibility economy.

Defining the possibility economy

The possibility economy is a social and economic system in which value arises from expanding the range of actions that people and communities can imagine, evaluate, and realize. Its products include commercial technologies and businesses, but its scope extends beyond markets. A public institution, scientific hypothesis, cultural work, local care system, coöperative, educational practice, or political reform can create value by giving people a capacity they did not possess before.

The term is intended to provide an alternative to the language of innovation, which has become managerial language. Innovation now tends to mean patents, startups, productivity gains, technology procurement, or the adoption of products developed elsewhere. Schools purchase digital platforms and describe themselves as innovative. Governments publish AI strategies and use technology adoption as evidence of progress. Companies automate work and present the result as innovation, even when the change removes worker authority, transfers public knowledge into private systems, or creates dependency on a vendor.

This use of innovation directs attention toward the recognized outcome. It gives less attention to the process through which people examine existing conditions, formulate questions, combine knowledge, develop alternatives, judge risks, organize resources, and secure the authority to act. Possibility describes this earlier and more open stage. It begins when a person or community can conceive of something worth bringing into the world and has a credible route for pursuing it.

The economic scale of this transition is substantial. UN Trade and Development estimates that the global AI market could grow from $189 billion in 2023 to $4.8 trillion in 2033. AI would account for 29% of the global frontier-technology market under that projection (UNCTAD, 2025). The International Monetary Fund estimates that AI will affect close to 40% of employment worldwide. Exposure rises to about 60% in advanced economies, compared with 40% in emerging-market economies and 26% in low-income countries (Cazzaniga et al., 2024).

Exposure does not mean that every affected job will disappear. AI can automate some tasks, improve others, and alter how expertise develops. Research has found measurable productivity gains in specific forms of knowledge work. In an experiment involving professional writing tasks, participants using ChatGPT completed their work about 40% faster, while independent evaluators rated the output 18% higher in quality (Noy & Zhang, 2023). A study of 5,179 customer-support agents found a 14% average increase in issues resolved per hour, with gains of 34% among novice and lower-skilled workers (Brynjolfsson et al., 2023).

These benefits depend on the task and the user’s judgment. In an experiment involving 758 consultants, participants using GPT-4 completed 12.2% more tasks and worked 25.1% faster on tasks that fell within the system’s capabilities. On a task outside that capability boundary, AI users were 19% less likely to produce the correct result (Dell’Acqua et al., 2025). AI can extend human capacity, but it can also produce confident errors and encourage users to stop examining the problem.

These findings help distinguish the possibility economy from a simple economy of automation. Automation seeks to complete existing tasks with less labor. Possibility involves using technology, knowledge, and coöperation to develop new purposes, services, institutions, and forms of action. AI can support both. The outcome depends on who controls its use and how the gains are distributed.

From data to possibility

The relationships among data, information, knowledge, possibility, and innovation provide a structure for understanding the educational problem.

Data consists of observations, records, measurements, signals, and representations. Data does not explain itself. People and institutions decide which phenomena to record, how to measure them, and which categories to apply.

Information develops when data is organized, compared, classified, or interpreted. A collection of measurements becomes informative through the questions and purposes applied to it.

Knowledge develops when people connect information with experience, context, memory, theory, practice, and judgment. Knowledge therefore involves more than possession of information. It includes an ability to interpret information within a situation and use it in relation to a purpose.

Possibility develops when people use knowledge to conceive of something that could exist. Its central question is not limited to predicting what will happen. It asks what a person or community wants to bring into the world and how existing knowledge could support that creation.

Innovation develops when a possibility becomes a practice, institution, product, policy, method, or relationship that expands human capacity. The result produces new data, alters the surrounding conditions, and generates further knowledge. The process forms a network of relationships and feedback loops rather than a hierarchy.

The familiar data-information-knowledge-wisdom pyramid presents these concepts as a sequence of increasingly valuable objects. This model understates the interpretive and political work involved in moving among them. Data must be produced, information must be organized, and knowledge must be developed through relationships among people, tools, institutions, and environments. Each transition reflects choices and forms of power.

Schools tend to operate within a restricted section of this process. Governments, curriculum developers, publishers, universities, assessment organizations, and technology providers organize much of the information before students encounter it. Teachers present that material, and students demonstrate that they can recognize, explain, reproduce, or apply it under specified conditions.

Schools describe this process as knowledge acquisition, although many assessments measure the reproduction and application of organized information. Students complete research projects and creative assignments, but institutions retain control over most topics, formats, timelines, and evaluation criteria. The learner enters after much of the intellectual design has occurred.

Generative AI competes with students at this point because it can convert organized information into an acceptable response. It can generate the essay, explanation, summary, calculation, presentation, or code that an assignment requests. Schools encounter AI as an answer-producing system because answer production occupies such a large part of formal education.

The wider economy uses AI across a broader set of relationships. Researchers use it to compare literature, identify patterns, generate hypotheses, and model scenarios. Designers use it to construct and revise prototypes. Organizations use it to translate, code, analyze, simulate, and coördinate. Communities can use it to document local conditions, prepare proposals, create language resources, and communicate with institutions.

The economic significance of AI lies in this capacity to reduce the cost of moving from information to a preliminary model, proposal, analysis, or prototype. The educational significance lies in whether students learn to direct that capacity toward purposes they can explain and defend.

The persistence of industrial education

The comparison between schools and factories has become a cliché, and the historical relationship is more complicated than the comparison suggests. Public education expanded literacy, civic participation, professional access, and social mobility. It also absorbed principles associated with industrial management, including standardization, centralized planning, efficiency measures, and the division of complex work into measurable tasks.

Raymond Callahan’s study of school administration documented the influence of business-efficiency movements and scientific management on educational leadership in the early twentieth century (Callahan, 1962). Samuel Bowles and Herbert Gintis examined how schools reproduced workplace hierarchies and rewarded behaviors associated with compliance and externally directed work (Bowles & Gintis, 1976/2011). Harry Braverman’s analysis of industrial labor described a related separation between conception and execution, in which managers designed the production process while workers completed fragmented tasks within it (Braverman, 1974).

That separation remains relevant to education. Institutions establish the objectives, construct the curriculum, frame the problems, and divide learning into tasks. Students carry out those tasks and receive credentials that signal their ability to perform within the system.

The problem becomes visible when the economic system changes. Schools can prepare students for roles created by industries that later relocate, contract, automate, or disappear. The students receive occupational preparation, but the community may retain little control over investment, ownership, strategy, or the creation of new institutions. Education can help people enter an existing economy without increasing their ability to shape the economy that follows.

Current workforce policy still emphasizes employer demand, occupational forecasts, and alignment between credentials and available positions. These concerns have practical importance, especially for learners who need a reliable route into employment. They become restrictive when schools define the current labor market as the outer limit of educational ambition.

The possibility economy requires people who can perform within existing systems and develop alternatives when those systems no longer provide a viable future. This shift places greater emphasis on problem formulation, knowledge creation, institutional design, enterprise formation, public action, and collective agency.

Possibility as an educational capacity

Possibility does not arise from knowledge alone. A person may understand a problem and see no feasible response. A community may possess extensive local knowledge while lacking capital, political authority, technical infrastructure, or access to institutional networks. Students may recognize a harmful condition while assuming that serious decisions belong to employers, officials, experts, or technology companies.

Possibility connects knowledge with agency. It develops when people can formulate alternatives and regard themselves as capable of pursuing them.

Anna Craft and her colleagues used the term possibility thinking to describe movement from existing conditions toward imagined alternatives. Their research connected possibility thinking with question formation, imagination, play, experimentation, and learner self-determination (Craft et al., 2013). The possibility economy extends this educational concept into questions of political economy, ownership, and institutional power.

The question “What do you want to bring into the world?” requires more than an expression of preference. Learners must explain the purpose of the proposed creation, examine the conditions surrounding it, identify the people affected, study previous efforts, gather evidence, and consider the resources required for implementation. They must assess potential harms, conflicts of interest, ownership, consent, and long-term responsibility.

This capacity is distributed unevenly. The OECD’s first international assessment of creative thinking, conducted as part of PISA 2022, examined students in 64 countries and economies. In 21 of those systems, more than half of students did not reach the baseline level of creative-thinking proficiency. Across OECD countries, students from advantaged socioeconomic backgrounds scored about 9.5 points higher than disadvantaged students on a 60-point scale (OECD, 2024). The ability to generate, evaluate, and improve ideas reflects opportunities that schools and families distribute in unequal ways.

Some students receive repeated opportunities to formulate questions, work on extended projects, consult mentors, recover from unsuccessful attempts, and present their ideas to influential audiences. Other students complete short tasks with known outcomes throughout most of their education. The difference affects who learns to see themselves as a creator of scientific, economic, institutional, or civic change.

Self-efficacy helps explain this pattern. Bandura (2000) argued that people act when they believe they can influence outcomes, and that collective efficacy increases the capacity of groups to pursue shared goals. Appadurai’s (2004) capacity to aspire places aspiration within cultural and institutional conditions. People develop an ability to pursue possible futures through practice, social support, access to knowledge, and experience with consequential decisions.

A possibility-oriented education would provide these experiences as part of the curriculum rather than reserve them for selective programs, well-resourced schools, or extracurricular activities.

Why predictable answers persist

Schools retain predetermined questions and outcomes because their accountability systems reward predictability. Students face grades, admissions decisions, financial pressures, and credential requirements. Teachers work within curriculum mandates, pacing expectations, evaluation systems, and public scrutiny. Administrators manage budgets, regulations, enrollment, test results, and institutional reputation. Policymakers seek comparable evidence of performance across classrooms, schools, and jurisdictions.

Open-ended work creates difficulty for these systems. Learner-generated questions complicate curriculum alignment. Community projects introduce conditions that teachers cannot control. Experiments produce incomplete results. Assessment requires professional judgment rather than automatic scoring.

Standardized tasks reduce these difficulties because they produce comparable outputs. Over time, the needs of measurement begin to shape the form of learning. Schools favor activities that can be scored, aggregated, and reported, even when those activities capture a narrow range of intellectual capacities.

AI can reinforce this preference. Automated assessment, predictive analytics, risk classification, surveillance, and intervention systems promise consistency and administrative control. They can process large quantities of student data while giving greater authority to categories, models, and thresholds that students and teachers may not be able to inspect.

AI can also support a different form of education. Students can use it to model alternatives, construct prototypes, compare sources, translate materials, test assumptions, and revise their work. This use requires schools to permit outcomes that teachers and platforms have not specified in advance.

This approach does not require a celebration of failure or the transfer of startup culture into schools. Students already experience unequal levels of economic and social risk. Educational experimentation should provide conditions in which learners can revise their work without allowing an unsuccessful attempt to determine their future. Assessment can examine the quality of inquiry, evidence, judgment, collaboration, revision, and responsibility.

The unequal structure of the possibility economy

The possibility economy is developing within severe inequalities in education, infrastructure, research capacity, and political influence.

In 2024, 5.5 billion people used the internet, while 2.6 billion remained offline. Internet use reached 93% of the population in high-income countries and 27% in low-income countries. The gap between urban and rural populations was also large, with 83% of urban residents online compared with 48% of rural residents (International Telecommunication Union, 2024).

Educational access presents another constraint. UNESCO reported that 251 million children and young people remained out of school in 2024. In low-income countries, 33% of the school-age population was out of school, compared with 3% in high-income countries (UNESCO, 2024). An argument about the creative and productive potential of AI must account for the hundreds of millions of people who lack reliable access to education and the billions who lack meaningful digital connectivity.

Even where access exists, control over AI remains concentrated. UNCTAD reported that 100 companies accounted for more than 40% of global business investment in research and development in 2022. About half of these companies were headquartered in the United States, and 13% were headquartered in China. No other developing country was represented among the largest 100 corporate research and development investors. The United States and China accounted for about one-third of AI publications and 60% of AI patents. The United States also held more than half of global computing capacity (UNCTAD, 2025).

This concentration affects which languages receive support, which problems attract investment, which datasets enter model development, and which organizations control the economic returns. Developing countries may gain access to AI services while remaining dependent on infrastructure, standards, and products owned elsewhere.

Coloniality operates through these relationships. Communities can contribute data, cultural materials, language, and labor without retaining authority over how these resources are used. AI systems trained through dominant institutions can treat external classifications and interpretations as universal. Local knowledge may remain absent because it has not been digitized, appears in a low-resource language, circulates through oral traditions, or belongs to communities that restrict its use.

Couldry and Mejias (2019) describe data colonialism as the appropriation of human life through data extraction. Birhane (2020) applies a related critique to algorithmic systems in Africa, where imported technologies can reproduce social and epistemic dependencies. UNESCO’s work on Indigenous data sovereignty emphasizes self-determination, consent, community governance, and the authority of Indigenous peoples over the use of their language, knowledge, images, and representations in AI systems (UNESCO, 2023).

Access to AI therefore represents one condition of participation. Sovereignty concerns the terms of that participation.

Technological sovereignty does not require every country to build a frontier model or own each part of the technical infrastructure. Most countries lack the capital, energy, and specialized labor required for that strategy. Sovereignty can develop through local-language resources, public-interest data governance, open standards, regional coöperation, procurement capacity, model evaluation, teacher expertise, curricular authority, and the legal capacity to reject systems that create unacceptable dependencies.

Educational sovereignty concerns the ability of a community to carry knowledge from lived experience into action under conditions it can influence.

GEFRI and the geography of possibility

The Global Education Futures Readiness Index provides one way to examine the conditions under which societies may participate in the possibility economy. GEFRI measures 217 countries and jurisdictions across infrastructure, human capital, school access and gender parity, innovation, and governance. The index does not measure the possibility economy or technological sovereignty directly, but it helps identify structural conditions that support or constrain the movement from knowledge to action.

The July 2026 update produced a global average score of 59.78 out of 100 and a median of 63.56. Ninety-five countries remained below the index’s progressing-readiness threshold of 60 points. Those countries contained 2.35 billion people. Only Denmark and Sweden reached the 90-point readiness target (Education Futures, 2026).

The differences among the five dimensions provide more useful evidence than the composite score alone. Infrastructure exceeded innovation by a substantial margin in 83.4% of countries. This pattern indicates that many societies are developing the capacity to connect to digital systems faster than they are developing the research, institutional, and creative capacity needed to produce and adapt technologies.

This imbalance has direct relevance to the possibility economy. Connectivity can expand access to tools created elsewhere without increasing local authority over their design or use. A country may possess strong digital infrastructure but limited research capacity, weak governance, or low tertiary participation. Under those conditions, AI adoption may increase consumption and dependency faster than it increases local creation.

[OPTIONAL GEFRI MAP OR COUNTRY PROFILE EMBED]

GEFRI also shows why the digital divide cannot be reduced to devices and broadband. Infrastructure supports participation, but human capital, access, innovation capacity, and governance influence whether people can convert technological access into sustained local action.

[KNOWLEDGE MAP PLACEHOLDER]

The interactive knowledge map will examine the relationships among data, information, knowledge, possibility, innovation, and the conditions that connect them. It will show how data becomes information through selection and classification, how information becomes knowledge through context and experience, and how knowledge can lead toward possibility when learners possess imagination, self-efficacy, agency, resources, and authority.

The map will also show the narrower pathway common in formal schooling, where organized information leads to assigned questions, acceptable responses, assessment, scores, and credentials. Generative AI will appear across both pathways. It can automate response production within the school pathway, and it can support research, modeling, prototyping, and experimentation within the possibility pathway.

Additional relationships will connect access, infrastructure, ownership, coloniality, sovereignty, governance, fear, assessment, collective agency, and institutional change. These connections will make visible the points at which possibility expands, stalls, or becomes controlled by another actor.

[EMBED KNOWLEDGE MAP HERE]

Education for the possibility economy

Education for the possibility economy would retain serious engagement with established knowledge. Learners cannot develop useful possibilities without understanding history, science, culture, institutions, and previous attempts to address similar conditions. The purpose of learning would extend beyond reproducing this knowledge in forms prescribed by the school.

Students would examine how data was gathered, how information was organized, and how interpretations change across social and cultural contexts. They would formulate questions, conduct research, identify uncertainty, develop alternatives, and examine the probable effects of their decisions. They would work with communities as participants in knowledge production and as parties with legitimate interests in the results.

The question “What do you want to bring into the world?” could organize this work across subjects and age levels. A student would need to describe the proposed contribution, the conditions that gave rise to it, the people who might use or experience it, and the evidence supporting its development. The student would need to identify the resources, relationships, permissions, and forms of expertise required to sustain it. Questions of ownership, consent, risk, and accountability would form part of the work rather than appear as an ethical supplement.

Assessment would examine reasoning, evidence, judgment, collaboration, revision, and contribution. Some projects would produce a durable result. Others would produce a stronger account of the problem, a more precise question, or evidence that a proposed approach should be discontinued. These outcomes can demonstrate learning when students document their decisions and respond to evidence.

Generative AI can support this work through source comparison, data analysis, translation, modeling, scenario development, prototyping, and communication. Teachers would help learners evaluate the reliability of outputs, identify missing perspectives, understand the provenance of data, and examine the consequences of using the system.

This model also extends Knowmad Society. The knowmad moves knowledge across contexts, develops networks, adapts to change, and creates value. In the possibility economy, mobility must be joined with the power to influence the conditions under which knowledge moves. Otherwise, flexibility can leave people adapting to priorities established by employers, platforms, and institutions they cannot influence.

Manifesto 25 and positive rebellion contribute a related principle. Education should give learners experience in examining inherited systems, questioning imposed limits, and developing credible alternatives with others. This requires knowledge, imagination, confidence, coöperation, and the capacity to organize action.

AI has exposed the limitations of education organized around expected answers. The economic effects will reach a large share of the global workforce, while access to the technology and control over its development will remain unequal. Schools can use AI to increase the efficiency of existing practices, or they can reconsider the relationship among knowledge, agency, and creation.

The possibility economy asks education to prepare people who can formulate questions, combine forms of knowledge, develop alternatives, and create new capacities with others. Its defining educational question is direct:

What do you want to bring into the world?

References

Appadurai, A. (2004). The capacity to aspire: Culture and the terms of recognition. In V. Rao and M. Walton (Eds.), Culture and public action (pp. 59–84). Stanford University Press. World Bank open-access edition

Bandura, A. (2000). Exercise of human agency through collective efficacy. Current Directions in Psychological Science, 9(3), 75–78. https://doi.org/10.1111/1467-8721.00064

Birhane, A. (2020). Algorithmic colonization of Africa. SCRIPTed, 17(2), 389–409. https://doi.org/10.2966/scrip.170220.389

Bowles, S., and Gintis, H. (2011). Schooling in capitalist America: Educational reform and the contradictions of economic life. Haymarket Books. Original work published 1976.

Braverman, H. (1974). Labor and monopoly capital: The degradation of work in the twentieth century. Monthly Review Press.

Brynjolfsson, E., Li, D., and Raymond, L. R. (2023). Generative AI at work (NBER Working Paper No. 31161). National Bureau of Economic Research. https://doi.org/10.3386/w31161 citeturn401868search1

Callahan, R. E. (1962). Education and the cult of efficiency: A study of the social forces that have shaped the administration of the public schools. University of Chicago Press.

Cazzaniga, M., Jaumotte, F., Li, L., Melina, G., Panton, A. J., Pizzinelli, C., Rockall, E. J., and Tavares, M. M. (2024). Gen-AI: Artificial intelligence and the future of work (Staff Discussion Note No. 2024/001). International Monetary Fund. https://doi.org/10.5089/9798400262548.006 citeturn650119search0

Couldry, N., and Mejias, U. A. (2019). The costs of connection: How data is colonizing human life and appropriating it for capitalism. Stanford University Press. Publisher page

Craft, A., Cremin, T., Burnard, P., Dragovic, T., and Chappell, K. (2013). Possibility thinking: Culminative studies of an evidence-based concept driving creativity? Education 3–13, 41(5), 538–556. https://doi.org/10.1080/03004279.2012.656671

Dell’Acqua, F., McFowland, E., III, Mollick, E., Lifshitz-Assaf, H., Kellogg, K. C., Rajendran, S., Krayer, L., Candelon, F., and Lakhani, K. R. (2025). Navigating the jagged technological frontier: Field experimental evidence of the effects of artificial intelligence on knowledge worker productivity and quality. Organization Science. https://doi.org/10.1287/orsc.2025.21838 citeturn401868search2turn401868search5

Education Futures. (2026, July 16). GEFRI mid-2026 update released: New data, persistent readiness gaps. https://educationfutures.com/post/the-gefri-mid-2026-update-new-data-persistent-readiness-gaps/ citeturn884602search4

International Telecommunication Union. (2024). Facts and figures 2024. https://www.itu.int/itu-d/reports/statistics/2024/11/10/ff24-internet-use/ citeturn650119search1turn650119search5

Noy, S., and Zhang, W. (2023). Experimental evidence on the productivity effects of generative artificial intelligence. Science, 381(6654), 187–192. https://doi.org/10.1126/science.adh2586 citeturn797744search0turn797744search3

Organisation for Economic Co-operation and Development. (2024). PISA 2022 results, Volume III: Creative minds, creative schools. OECD Publishing. https://doi.org/10.1787/765ee8c2-en citeturn895364view3turn437820search6

Organisation for Economic Co-operation and Development. (2025). Results from TALIS 2024: The state of teaching. OECD Publishing. https://doi.org/10.1787/90df6235-en citeturn895364view2turn975613search12

UN Trade and Development. (2025). Technology and innovation report 2025: Inclusive artificial intelligence for development. United Nations. https://unctad.org/publication/technology-and-innovation-report-2025 citeturn650119search2turn450782search7

UNESCO. (2023). Indigenous people-centered artificial intelligence: Perspectives from Latin America and the Caribbean. UNESCO publication page

UNESCO. (2024, October 31). 251 million children and youth still out of school, despite decades of progress. UNESCO press release citeturn650119search7