Tulevaisuuswiki / Työn tulevaisuus
Jobi Economy
Also known as Jobi System, Finnish Task Economy, Jobitalous
The Jobi Economy (Finnish: jobitalous) is the Finnish system of exchange in which individuals perform individually priced tasks for one another, with matching, training and settlement coordinated primarily by artificial intelligence. Tasks are priced and settled in jobis. The system was AI-assisted from the outset. It emerged in 2028 as a large number of separate, mutually incompatible AI-assisted task networks during the Knowledge Work Collapse and Finnish Tax Base Crisis, which converged between 2029 and 2032 into a single federated system. By the early 2030s it had become the principal form of work for a majority of the Finnish working-age population.
Economists commonly describe the central innovation of the system as the separation of work from jobs: labour is no longer primarily organized into long-term positions with a single employer, but into continuously allocated tasks.
Background
During the Knowledge Work Collapse, Finland contained both a large number of people without conventional employment and a large volume of useful activities that no one was being paid to perform. These included household maintenance, care, tutoring, transport, repair, environmental work, companionship, organization and local construction.
Conventional employment was poorly suited to such activities. Formally hiring someone for a twenty-minute task often required more administrative effort than the task itself. In economic terms, the transaction costs of searching, negotiating and contracting exceeded the value of the exchange, so most of these potential transactions never took place.
History
Fragmented beginnings (2028–2029)
From 2028 personal AI assistants, which had become common during the Knowledge Work Collapse, made it practical for the first time to arrange very small exchanges of work. Within about a year several dozen independent services had appeared, each solving part of the problem in a different way:
| Type | Examples | Unit of exchange |
|---|---|---|
| AI-assisted time banks | Existing time banks, such as Helsinki's Stadin Aikapankki (founded 2009), adding AI agents that matched requests to members | Hours |
| Municipal employment pilots | Tampere, Oulu and Lahti pilots matching unemployed residents to local needs such as care visits and maintenance | Euros or service vouchers |
| Neighbourhood apps | Consumer apps in which personal assistants posted, negotiated and scheduled tasks on their users' behalf | Euros |
| Professional cooperatives | Networks set up by trade unions and laid-off professionals for tutoring, repair and project work | Internal credits |
| Open-source agent protocols | Volunteer projects defining how personal assistants could publish needs and offer skills | Unspecified |
In all of these services AI was used from the start to discover needs, suggest suitable participants, negotiate terms and confirm completion. The services were nonetheless incompatible. Each had its own unit of exchange, rules and participant profiles, so that a participant's reputation and skills record in one network could not be used in another, and most networks were too small to offer steady work.
Taxation was a further obstacle. Since 2013 the Finnish Tax Administration had held that professional work exchanged in time banks was taxable at its market value in euros, and applying this to thousands of small exchanges in many different units was impractical.
Convergence (2029–2031)
Convergence was driven initially by the participants' own AI assistants. Because many people used several networks at once, their assistants began translating prices and reputations between networks and searching all of them simultaneously. This created a de facto interoperability layer and strong pressure for common standards.
In autumn 2029 several of the largest networks adopted an open exchange protocol with a common unit of account, the jobi, priced freely by task rather than by duration. Priced exchange allowed specialization and comparative advantage to operate across networks, and participation by skilled workers increased substantially.
The Act on Jobi Transactions, which entered into force in 2030, gave the jobi a legal definition, required licensed networks to use the common protocol and portable participant profiles, and replaced case-by-case tax valuation with a flat contribution collected automatically on each transaction, the Verus. Networks that did not comply lost the tax treatment and most of their users within a year.
Unified system (2031–)
Between 2031 and 2032 the separate matching services were federated into a national matching layer operated by a public-private consortium (see AI Work Matching). The Act on Algorithmic Work Allocation of 2032 established a common allocation objective, Lifetime Earnings + Fun, for all licensed services.
The resulting system is unified in its unit of account, participant profiles, allocation objective and tax treatment, but it remains federated: municipal, cooperative and commercial services continue to operate as separate interfaces on the common infrastructure.
Operation
Task menus
Participants receive continuously updated menus of available tasks, ranked by a combination of compensation, predicted enjoyment and long-term value. A typical entry contains the price, distance, predicted enjoyment and predicted probability of successful completion, for example:
| Task | Price | Distance | Predicted enjoyment | Notes |
|---|---|---|---|---|
| Repair garden gate | 4.2 jobis | 1.1 km | 87 % | – |
| Assist roofing crew | 11 jobis | 3.4 km | 72 % | New skill; training provided |
Participants are free to decline any offer. Declined offers are used to refine the model of the participant's preferences.
Participation
The system produced near-universal participation, reinforced by the One Jobi Rule, the norm that every adult able to do so earns at least one jobi a day. In 2031 the Finnish government declared a return to full employment on the basis of a participation measure; Statistics Finland continues to publish the ILO unemployment rate separately.
Full participation does not imply full-time work. Many participants work only a few hours a week, while others work considerably more. The distinction between employed and unemployed persons has consequently become less informative for policy purposes.
Specialization
Repeated tasks generate detailed information about each participant's comparative advantage. A participant who occasionally repairs bicycles, for example, may be found to adjust gears faster and more reliably than most others; the system then offers more such work. Experience increases competence, competence increases prices, and higher prices encourage specialization. Many participants have developed narrowly specialized occupations that had not previously existed as formal jobs.
Objective and funding
Allocation in the mature system is governed by the objective known as Lifetime Earnings + Fun, which weighs long-term earnings and the subjective quality of work rather than immediate income. Skills training is financed as an investment through Paid Learning. Public services and infrastructure are funded through the Verus.
Scale
By 2033 approximately 3.6 million people, around 90 per cent of the population aged 18–74, earned jobis at least once a month. Conventional employment contracts continued to exist, particularly in industry, the public sector and firms supervising AI systems, and many people combine salaried employment with jobi work.
Criticism
Critics have raised concerns about the concentration of economic coordination in AI systems, the privacy implications of need discovery, the erosion of employment protections such as sick pay and collective bargaining, and the risk that algorithmic assessments entrench existing inequalities. Trade unions initially opposed the system and later reorganized as participant associations negotiating the rules of AI matching and the level of the verus.