The Main Point

The traditional promise of economic growth was straightforward: businesses produced more, earned more, expanded and hired. Wages carried the benefits into households, consumption rose and prosperity spread through the economy. Artificial intelligence may break that chain at a critical point. A company can now increase output and profit while expanding headcount much more slowly-or not at all.

The American picture described by The Wall Street Journal combines strong corporate revenue and earnings with unusually weak job creation. The article does not attribute everything to AI: demographics, migration, interest rates and the economic cycle also matter. Yet the direction is increasingly visible. The historic link between output and employment is weakening, and AI can accelerate that separation.

Georgia is not yet in a jobless-growth economy

Georgia’s data present a more nuanced picture. According to Geostat’s preliminary estimate, real GDP grew 9.0% year on year in the first quarter of 2026. Over the same period, business-sector employment rose from 770,951 to 799,443, an increase of 3.7%. Jobs did not decline. But the 5.3-percentage-point gap between GDP growth and business-employment growth indicates that a significant part of additional output is already being created without a proportionate increase in headcount.

The two measures cannot be treated as identical: GDP covers the whole economy, while business-employment statistics cover enterprises. Still, the gap is a useful directional signal. Information and communication provides a sharper case. Its real output grew 36.0% in the first quarter of 2026, while business employment in the sector rose 7.3%—a gap of 28.7 percentage points. This is not an official productivity measure, but it shows the capacity of a digital sector to generate value far faster than it adds people.

The broader labour-market data also warrant caution. In the first quarter of 2026, the unemployment rate fell by 0.3 percentage points year on year to 14.4%, while labour-force participation declined by 1.4 points and the employment rate by 1.0 point. Even lower unemployment can coexist with people leaving the labour market. Rapid GDP growth does not necessarily mean that a larger share of citizens participates in that growth through work.

AI changes how firms expand before it eliminates jobs

AI’s first effect is often not a mass layoff. The more common shift is quieter: a company serves more customers with the same team; a vacancy is not opened; a departing employee is not replaced; entry-level tasks are absorbed by software; and one manager oversees a larger process with fewer people. Employment may not contract, but it no longer grows alongside revenue and output.

This distinction matters. Public debate watches announced layoffs, while AI’s largest labour effect may be jobs that are never created. A 20% expansion that once required additional operators, analysts or administrative staff can now be handled partly by existing employees using AI. The statistics do not look like a crisis: the company grows, current employees remain and unemployment does not jump. But the entry route for the next generation narrows.

Who receives the productivity dividend

AI’s economic importance is not determined only by the hours it saves. The decisive issue is ownership of the value saved. When the same revenue is produced with lower labour input, the gain can flow through four channels: higher profit for owners; higher pay or shorter hours for workers; lower prices or better service for consumers; and additional tax revenue that finances public goods.

None of these outcomes is automatic. Without competition, the gain may stay in margins. Without worker bargaining power, employees may simply face greater work intensity. When a few technology providers dominate, part of the value may leave a local business as fees paid to foreign platforms. AI-led growth and broad social welfare are therefore not the same metric.

Georgia’s four distribution channels

Profit and ownership

When AI allows a firm to scale without rapidly expanding headcount, capital owners receive the first benefit. This matters in Georgia because most households participate in growth through wages rather than through shares, pension assets or business ownership. Higher profits do not automatically reach families that own no stake in productive companies.

Pay and working time

In the first quarter of 2026, average monthly remuneration in Georgia’s business sector rose from GEL 2,169.6 to GEL 2,335.9-about 7.7%. This is meaningful, but it is a nominal average and does not show whether workers received the full productivity gain. Future monitoring should include median pay, hours, bonuses and disparities between sectors, not only the average wage.

Prices and consumers

Lower AI-enabled costs reach consumers mainly through competition. If firms reduce prices or improve quality, productivity is broadly distributed. In a concentrated market, automation may remain almost invisible to the customer while margins rise. Competition policy therefore becomes part of AI social policy.

The state and taxation

If payroll grows more slowly than GDP and profit, a tax system heavily dependent on labour income captures less of the productivity dividend. The answer is not a mechanical ‘robot tax’. More important are effective profit taxation, transparency in digital activity, competition and social investment that helps people acquire new productive roles.

The first job may be more vulnerable than the old job

Generative AI most readily affects tasks involving drafting, data entry, standard analysis, first-pass coding and routine customer replies. These are often the tasks assigned to junior employees. The ILO’s 2025 index estimates that roughly one quarter of global employment is in occupations with some exposure to generative AI, while 3.3% falls into the highest exposure category. The ILO stresses that transformation, rather than complete replacement, is the most likely outcome.

Transformation can still be unequal. Senior professionals become more productive with AI, while juniors lose the basic tasks through which experience is built. If organisations pursue cost reduction alone, they may later discover that they have broken the pipeline that produces future senior talent. Good AI management should redesign junior roles, not merely remove them.

Georgia could split into two economies

One economy would consist of a small group of highly productive companies that use AI, sell internationally and generate substantial value per employee. The other would retain low-productivity services, self-employment and activities with limited access to technological capital. Wages and profits would rise quickly in the first; in the second, people would compete both with each other and with firms operating at a radically lower cost.

The divide may also be geographic. Technology, finance and professional services concentrated in Tbilisi are likely to receive the gains first, while diffusion to smaller cities and rural areas may be slower. If advanced AI remains a tool of a few large companies, national productivity can rise together with regional and income inequality.

How to distribute the AI dividend

The first task is measurement. Alongside GDP, Georgia should regularly track real value added per worker, labour’s income share, median pay, working hours, vacancies and the number of entry-level positions. Unemployment alone will reveal the change too late.

The second task is distribution within the firm. Productivity gains can be linked to performance bonuses, profit sharing, employee equity, additional leave or shorter working weeks. This is not simply a social gesture. When employees see AI only as a threat, they hide knowledge and resist implementation; when they share in the gains, adoption becomes a joint project.

The third task is to build new entry routes. Government and business will need internships, occupational-transition programmes and AI training tied to real workflows rather than generic courses. Workers must learn not only to prompt a model, but to verify output, handle clients, accept responsibility and make decisions that cannot be automated.

The fourth task is technological access for small and medium-sized firms. If advanced AI is available only to large companies, the productivity gap will reinforce market concentration. Shared infrastructure, secure Georgian-language tools, sectoral data and practical advisory support can allow smaller firms to join the growth as producers rather than mere customers.

A new social contract for growth

BTU researchers conclude that Georgia has not yet reached a point where the economy grows and jobs cease to be created. First-quarter 2026 business statistics still show employment growth. But the different speeds of GDP, digital-sector output and employment already pose a question that will become more urgent as AI spreads: is more production sufficient, or does the country need new forms of economic participation?

The old model distributed growth mainly through job creation. That channel may weaken. Georgia will therefore need a broader distribution system: higher pay, participation in profit, wider ownership, competitive prices, accessible technology and continuous occupational transition. AI’s success should not ultimately be judged by GDP growth alone. The better measure is how many people gain stronger income, greater choice and a dignified place in the economy.

Sources

• The Wall Street Journal – Greg Ip, ‘The Jobless Economic Boom Has Arrived’, 14 აგვისტო 2026. User-provided issue; no direct public URL was supplied.
• National Statistics Office of Georgia – Gross Domestic Product – https://www.geostat.ge/en/modules/categories/23/gross-domestic-product-gdp
• National Statistics Office of Georgia – Business Statistics – https://www.geostat.ge/en/modules/categories/195/business-statistics
• National Statistics Office of Georgia – Employment and Unemployment – https://www.geostat.ge/index.php/en/modules/categories/683/Employment-Unemployment
• International Labour Organization – Generative AI and Jobs: A Refined Global Index of Occupational Exposure – https://www.ilo.org/publications/generative-ai-and-jobs-refined-global-index-occupational-exposure
• BTU research and analytical papers on the agentic economy, labour and distribution of economic gains. No public URL is available.