Separate five questions
What tasks can the technology perform?
Where is it actually deployed?
Does real output per unit of labor, capital, or time improve?
Which tasks, occupations, hours, skills, and jobs change?
Who receives wages, prices, profits, ownership, public revenue, services, or time?
ILO's 2025 occupational-exposure work similarly cautions against treating GenAI exposure as automatic job disappearance.
Productivity has multiple legitimate claimants
Higher wages, lower prices, profit, investment, ownership returns, public revenue, better quality, more output, new products, and shorter hours can all be real gains. Human Scale's political value is not that every gain becomes leisure. It is that human time and broadly shared material gains should be legitimate outcomes of progress, not accidental leftovers.
Ownership is one mechanism, not the whole answer
Can align workers with enterprise upside and voice, but can also concentrate job and savings risk in the same firm.
Can combine ownership and governance, but require viable capital, management, market position, and member coordination.
Can socialize returns or protect public goods, but needs governance, transparency, and correction of poor performance.
Pensions, retirement accounts, diversified funds, and household assets can spread capital returns without firm-level governance.
Competition distributes gains too
When entry and switching are realistic, productivity can reach consumers as lower prices, better quality, and new services. When markets are concentrated or locked in, firms may retain more of the gains. Interoperability, procurement, entry barriers, labor mobility, and market structure therefore matter alongside taxes and transfers.
Worker voice matters during implementation
People affected by workplace AI can possess important operational knowledge and legitimate interests in how monitoring, scheduling, job design, evaluation, and automation change. ILO case studies document social dialogue as one real implementation mechanism.
Human Scale remains plural about institutional form: direct voice, unions, works councils, ownership, standards, grievance systems, and other mechanisms can compete on outcomes.
Material security is transition infrastructure
A worker with savings, portable benefits, realistic jobs, housing stability, transportation, childcare, and accessible training experiences technological transition differently from someone one paycheck from crisis. Security changes the practical ability to say no, learn, move, start something, or wait for a better match.
Retraining is not a ritual answer
Training works when real labor demand exists, the skill is learnable, credentials matter, people can afford the transition, and the resulting jobs are accessible. Measure downstream capability, earnings, employment, and user outcomes — not certificates issued.
Shorter work is one distribution channel
Saved task time can become more output, fewer workers, higher quality, faster service, lower prices, profit, training, or shorter schedules. Human Scale keeps shorter work as an experiment class, not an automatic entitlement created by a faster model.
Doctrine-wide corrections
Institutions shape the social result.
Employee, cooperative, public, household, and diversified ownership solve different problems.
Lower prices and better services can be distribution too.
Transitions are different when people can realistically refuse bad options.
Do not count enrollment as successful adaptation.
Shorter work remains one possible productivity dividend.
What would change our minds?
Reduce ownership emphasis where it produces little wealth or agency, concentrates risk, or is inferior to wages, competition, public services, or other mechanisms. Strengthen specific ownership forms where strong evidence shows durable gains without unacceptable risk. Reduce shorter-work emphasis where it lowers security or service or intensifies work. Revise the whole framework if real-world AI gains arrive through channels Human Scale has underweighted.