AI is changing work more quickly than aggregate employment. The strongest current evidence points to task reconfiguration and role redesign before mass job elimination: roughly one in four workers globally are in occupations with some GenAI exposure, while firms with higher AI exposure are also showing faster productivity, wage and in some studies headcount growth. The strategic issue is therefore not “how many jobs disappear?” but which tasks move to machines, which human capabilities become more valuable, and whether organisations redesign roles, management and learning fast enough to capture productivity without hollowing out expertise.
Research method
This report triangulates current primary/public-sector evidence, large-sample research and relevant industry/platform documentation. It prioritises observed behaviour over forecasts, distinguishes vendor claims from neutral statistics, and uses an explicit 8i judgment layer only where source evidence ends. The analysis is current through 15 September 2026.
1. The evidence base: transformation before replacement
The ILO’s refined 2025 global exposure index covers nearly 30,000 tasks and estimates that one in four workers worldwide are in occupations with some degree of GenAI exposure, but only 3.3% of global employment sits in the highest exposure category. The ILO’s central conclusion is that transformation is more likely than wholesale redundancy because most exposed occupations still require substantial human input. S1
Exposure is very uneven. High-income economies have a much larger share of employment in exposed knowledge and clerical work than low-income economies, and the ILO identifies a significant gender asymmetry because clerical occupations employ many women. S1 This means workforce strategy cannot rely on a single global displacement rate.
2. What is already changing in labour markets
PwC’s 2026 AI Jobs Barometer, based on more than one billion job advertisements across six continents, finds skills changing more than twice as quickly in highly AI-exposed jobs as in less-exposed jobs. Jobs requiring explicit AI skills are growing much faster than the total jobs market and attract a large wage premium. S2
Stanford’s 2026 AI Index finds labour-market effects are uneven and most visible in hiring pipelines and younger workers in exposed occupations, while broad employment losses have not yet appeared at economy-wide scale. S3 The correct interpretation is early structural pressure, not proof of universal displacement.
3. The emerging division of labour
The strongest pattern is a split between machine execution and human direction. Microsoft’s 2026 Work Trend Index finds organisational factors such as culture, manager support and talent practices account for substantially more variation in reported AI impact than individual attitudes alone. S4 This supports an operating-model view of work transformation.
A practical division of labour is emerging: AI prepares, searches, drafts, simulates, codes and executes bounded steps; humans frame the problem, set goals, judge trade-offs, manage exceptions, validate consequential outputs, own relationships and remain accountable for outcomes. This is not a fixed boundary: the line moves as reliability and controls improve.
4. Roles are broadening, not simply shrinking
PwC describes two labour-market paths: ‘professionalised’ roles, where AI acts as a force multiplier for experts, and ‘democratised’ roles, where AI lowers entry barriers. Professionalised roles are growing faster and showing stronger wage growth in its data. S2
This creates role convergence. Employees can increasingly perform adjacent analytical, creative, technical or operational tasks with AI assistance, while deep specialists remain essential for high-risk validation. The organisation therefore needs fewer rigid task silos and more outcome-based roles, but it also needs protected centres of expertise to prevent capability decay.
Sources and notes
- ILO, Generative AI and Jobs: A Refined Global Index of Occupational Exposure (20 May 2025). [Source](https://www.ilo.org/publications/generative-ai-and-jobs-refined-global-index-occupational-exposure)) — Global task-level exposure analysis; 29,753-task methodology.
- PwC, 2026 Global AI Jobs Barometer (15 Jun 2026). [Source](https://www.pwc.com/gx/en/1/services/ai/ai-jobs-barometer.html)) — More than one billion job advertisements across six continents.
- Stanford HAI, 2026 AI Index — Economy. [Source](https://hai.stanford.edu/ai-index/2026-ai-index-report/economy)) — Cross-source synthesis on adoption, productivity and labour effects.
- Microsoft, 2026 Work Trend Index (5 May 2026). [Source](https://www.microsoft.com/en-us/worklab/work-trend-index/agents-human-agency-and-the-opportunity-for-every-organization)) — 20,000 respondents across 10 markets; vendor study, used for organisational-pattern evidence.
8i TECHNOLOGIES
AI and the Redesign of Work
Jobs, Skills, Roles, Management and the New Division of Labour Between Humans and AI
Reference date: 15 September 2026
See what matters next.
| Executive answer AI is changing work more quickly than aggregate employment. The strongest current evidence points to task reconfiguration and role redesign before mass job elimination: roughly one in four workers globally are in occupations with some GenAI exposure, while firms with higher AI exposure are also showing faster productivity, wage and in some studies headcount growth. The strategic issue is therefore not “how many jobs disappear?” but which tasks move to machines, which human capabilities become more valuable, and whether organisations redesign roles, management and learning fast enough to capture productivity without hollowing out expertise. |
|---|
Research method
This report triangulates current primary/public-sector evidence, large-sample research and relevant industry/platform documentation. It prioritises observed behaviour over forecasts, distinguishes vendor claims from neutral statistics, and uses an explicit 8i judgment layer only where source evidence ends. The analysis is current through 15 September 2026.
1. The evidence base: transformation before replacement
The ILO’s refined 2025 global exposure index covers nearly 30,000 tasks and estimates that one in four workers worldwide are in occupations with some degree of GenAI exposure, but only 3.3% of global employment sits in the highest exposure category. The ILO’s central conclusion is that transformation is more likely than wholesale redundancy because most exposed occupations still require substantial human input. [S1]
Exposure is very uneven. High-income economies have a much larger share of employment in exposed knowledge and clerical work than low-income economies, and the ILO identifies a significant gender asymmetry because clerical occupations employ many women. [S1] This means workforce strategy cannot rely on a single global displacement rate.
2. What is already changing in labour markets
PwC’s 2026 AI Jobs Barometer, based on more than one billion job advertisements across six continents, finds skills changing more than twice as quickly in highly AI-exposed jobs as in less-exposed jobs. Jobs requiring explicit AI skills are growing much faster than the total jobs market and attract a large wage premium. [S2]
Stanford’s 2026 AI Index finds labour-market effects are uneven and most visible in hiring pipelines and younger workers in exposed occupations, while broad employment losses have not yet appeared at economy-wide scale. [S3] The correct interpretation is early structural pressure, not proof of universal displacement.
3. The emerging division of labour
The strongest pattern is a split between machine execution and human direction. Microsoft’s 2026 Work Trend Index finds organisational factors such as culture, manager support and talent practices account for substantially more variation in reported AI impact than individual attitudes alone. [S4] This supports an operating-model view of work transformation.
A practical division of labour is emerging: AI prepares, searches, drafts, simulates, codes and executes bounded steps; humans frame the problem, set goals, judge trade-offs, manage exceptions, validate consequential outputs, own relationships and remain accountable for outcomes. This is not a fixed boundary: the line moves as reliability and controls improve.
4. Roles are broadening, not simply shrinking
PwC describes two labour-market paths: ‘professionalised’ roles, where AI acts as a force multiplier for experts, and ‘democratised’ roles, where AI lowers entry barriers. Professionalised roles are growing faster and showing stronger wage growth in its data. [S2]
This creates role convergence. Employees can increasingly perform adjacent analytical, creative, technical or operational tasks with AI assistance, while deep specialists remain essential for high-risk validation. The organisation therefore needs fewer rigid task silos and more outcome-based roles, but it also needs protected centres of expertise to prevent capability decay.
5. The entry-level paradox
AI removes some of the routine work through which junior employees traditionally learn. PwC finds AI-exposed entry-level roles are much more likely to demand skills historically associated with seniority, including judgement and leadership. [S2]
This creates a serious design challenge: if organisations automate apprenticeship tasks without replacing the learning pathway, they may optimise short-term throughput while degrading the future supply of experienced professionals. Workforce redesign must therefore include deliberate practice, supervised judgement and staged responsibility.
6. A maturity model for work redesign
| Stage | Operating pattern | Management implication |
|---|---|---|
| 1. Tool adoption | Individuals use copilots in existing roles. | Track quality and learning, not login counts. |
| 2. Task redesign | Repeatable tasks are reallocated between people and AI. | Clarify review, escalation and accountability. |
| 3. Role redesign | Job boundaries broaden or specialise around outcomes. | Rewrite role expectations, metrics and career paths. |
| 4. Workflow redesign | Cross-functional processes are rebuilt around AI execution. | Change decision rights, controls and handoffs. |
| 5. Organisational redesign | Teams, spans, layers and capability pools change. | Rebalance management, talent and investment systemically. |
7. What leaders should measure
- Accepted output per paid hour, not AI usage volume.
- Cycle time and rework, including human review burden.
- Quality, error and exception rates by task type.
- Time reallocated to higher-value work and whether that value is realised.
- Skill acquisition and skill decay, especially in early-career roles.
- Employee agency, workload and job quality—not only productivity.
- Internal mobility and the rate at which roles are recomposed rather than eliminated.
8. 8i assessment
The near-term competitive advantage is organisational learning speed. AI capability is diffusing rapidly; what remains scarce is the ability to redesign work without losing judgement, trust or future expertise. The most valuable 8i intelligence asset would therefore track task redistribution, role convergence, skill premiums, entry-level changes and realised productivity by sector—not headline job-loss forecasts.
Source register
Primary and high-value sources used. Time-sensitive claims were checked against the latest available release as of the reference date. Commercial/vendor research is identified as such and is not treated as neutral market truth without corroboration.
[S1] ILO, Generative AI and Jobs: A Refined Global Index of Occupational Exposure (20 May 2025). Source — Global task-level exposure analysis; 29,753-task methodology.
[S2] PwC, 2026 Global AI Jobs Barometer (15 Jun 2026). Source — More than one billion job advertisements across six continents.
[S3] Stanford HAI, 2026 AI Index — Economy. Source — Cross-source synthesis on adoption, productivity and labour effects.
[S4] Microsoft, 2026 Work Trend Index (5 May 2026). Source — 20,000 respondents across 10 markets; vendor study, used for organisational-pattern evidence.
[S5] OECD, AI and Work topic hub. Source — Policy synthesis on productivity, job quality, agency, bias and adoption.
[S6] ILO, Artificial intelligence adoption and its impact on jobs (31 May 2025). Source — G20-focused synthesis of employment and job-quality evidence.