AI adoption and the future of work: AI adoption produced a 5% net job loss but a 9.6% productivity gain across five AI-exposed sectors

Overview

This AlphaWise survey examines how AI is affecting employment, productivity, use cases, governance and implementation challenges across banks, software and services, tech hardware and equipment, semiconductors, and professional services in five countries. The report finds measurable efficiency gains alongside substantial workforce reallocation and expects productivity benefits to increase over the next year.

Core views

Morgan Stanley's second-wave AlphaWise survey covered 808 professionals with detailed knowledge of corporate AI strategy and implementation in the US, UK, Germany, Japan and Australia. The sample was limited to companies in five AI-exposed sectors that had implemented or developed AI solutions for at least 12 months and planned to continue over the following year. Across the sample, companies reported an average net loss of 5% of positions over the previous 12 months, suggesting that AI adoption has already translated into workforce restructuring among relatively advanced adopters. The gross workforce effects were much larger than the net figure: 12% of jobs were eliminated and another 15% of positions were not backfilled, for a combined 27%. These changes were partly offset by 22% new hires, while 25% of employees were retrained and 16% redeployed; only 11% of positions were unaffected. Japan reported the largest net job loss at 10%, followed by the UK at 6%, the US at 5% and Australia at 4%, whereas Germany was the only market reporting a net gain of 1%. Germany also had the lowest eliminated-or-not-backfilled share at 21%, the lowest elimination rate at 9%, and the highest retraining share at 27%. Sector outcomes varied substantially. Semiconductors had the highest net job loss at 8% and the highest combined elimination and non-backfill rate at 29%; the sector also led in redeployment, indicating that its workforce adjustment was not solely a reduction in headcount. Software and services had a 7% net loss, tech hardware 5%, professional services 4%, and banks the lowest at 3%. Banks nevertheless reported 25% of positions eliminated or not backfilled against 23% new hires, while retraining was particularly prominent in the sector. The report identifies uneven exposure across worker groups. Offshore employees were most affected by eliminated or non-backfilled positions, at 41%, followed by permanent employees at 38% and contract or temporary employees at 26%. The impact differed by sector: offshore roles were more exposed in banks and software and services, while permanent workers were more exposed in semiconductors. Employees with two to 10 years of experience accounted for the greatest share of eliminations and non-backfills, at 35% for each category; those with two to five years were especially affected, accounting for 18% of eliminations and 19% of non-backfilled roles. Yet this cohort was also central to adjustment, with employees having two to 10 years of experience accounting for 41% of hires, 39% of retraining and 40% of redeployment. Looking forward, 72% of companies expected to hire candidates with six to 10 years of experience and 65% expected to hire those with two to five years of experience, while graduate hiring could be around 40%. Reported productivity gains provide the counterweight to the workforce effects. Net productivity increased by an average 9.6% over the past 12 months. The UK led countries at 10.3%, followed by the US at 10.2%; Germany was lowest at 8.4%. Software and services led sectors at 10.4%, followed by banks at 10.0%, professional services at 9.6%, tech hardware at 8.6% and semiconductors at 8.2%. Small companies with fewer than 50 employees reported the highest gain at 11.3%, while firms generating under $1 million of 2025 revenue reported 13.3%, although the report flags low bases for some smaller revenue categories. IT/software development was the leading area of productivity improvement across countries, company sizes, revenue bands and AI-adoption maturities. Professional services differed somewhat, with customer service/support cited by 52%, ahead of IT/software development at 49% and legal and compliance at 48%. AI implementation is no longer nascent for this sample: companies had used AI solutions for an average 2.9 years, and 35% had done so for more than three years. Japan had the longest average tenure at 3.1 years and the US the shortest at 2.7 years. Banks, tech hardware and semiconductors averaged three years of implementation, compared with 2.8 years for professional services. Higher-revenue and larger companies generally had longer implementation histories. The report also documents concrete sector use cases: personalized banking and customer insights led in banking; search and software-development automation led in software and services; search and AI-enhanced business-intelligence dashboards led in tech hardware; semiconductor testing automation led in semiconductors; and search, automated content generation and automated summarization were prominent in professional services. Companies expect AI-related productivity gains to rise to 11.7% on average over the next 12 months. Japan had the highest country expectation at 12.3%, followed by the US at 12.1%; sector expectations ranged from 10.9% in tech hardware to 12.2% in professional services. Firms with only one year of AI implementation expected the largest next-year gain, at 15.7%. IT/software development was again expected to generate the largest productivity improvement across most groups, although lower-revenue companies expected marketing to lead. Offshore employees were expected to remain the group most likely to see eliminated or non-backfilled positions, while permanent employees were expected to be most exposed in the US, Japan, semiconductors and professional services. Implementation constraints and governance remain central to the report's interpretation of adoption. Data readiness and access, AI skills shortages, legacy-system integration, and trust, security, privacy and reputational risks were recurring challenges. Security, legacy integration, data readiness and skills shortages were all major issues for banks; semiconductor respondents placed particular emphasis on trust, security, privacy and reputational risk. Most respondents reported governance measures already in place: 93% had a responsible-AI policy, comprising 54% comprehensive and 39% focused policies. Across markets, 53% viewed people primarily as collaborative partners augmenting AI outputs with human judgment, while 33% preferred an oversight role focused on responsible use, compliance and risk management.

Analysis framework

The report uses a five-country corporate survey to compare reported AI-driven workforce outcomes, productivity gains, use cases, adoption maturity, governance and implementation barriers. It analyzes results by country, sector, company size, revenue, employee type, work experience and years of AI implementation, then contrasts the prior 12 months with expectations for the next 12 months.

Methodology notes

• Other

  AlphaWise corporate AI adoption survey

  Morgan Stanley surveyed 808 professionals in April 2026 whose companies had at least one year of AI implementation or development. The survey compares self-reported employment, productivity and adoption outcomes across countries and sectors.

• Corporate Fundamentals and FinanceEarnings Quality Analysis

  Business-outcome KPI tracking

  The report assesses AI outcomes through measures such as quality and accuracy, productivity and time saved, financial impact or ROI, and other operating KPIs, rather than relying on employment changes alone.

Key data

• Survey sample808 online interviewsConducted in April 2026 across five countries and five AI-exposed sectors
• Average net workforce change-5%Average net loss of positions over the prior 12 months
• Positions eliminated or not backfilled27%Partially offset by 22% new hires
• Past-12-month net productivity gain9.6%Average reported increase from AI implementation
• Expected next-12-month net productivity gain11.7%Average company expectation
• Japan net job loss10%Highest country-level net loss
• Semiconductor net job loss8%Highest sector-level net loss
• Responsible AI policy adoption93%54% comprehensive policies and 39% focused policies

Impact & implications

The report indicates that AI adoption is already changing workforce composition rather than simply reducing jobs: retraining, redeployment and selective hiring coexist with eliminations and non-backfills. It also suggests that productivity benefits are broadening and expected to accelerate, but their realization remains linked to data readiness, skills, legacy-system integration, security and governance.

Risks

• Data readiness and access can constrain AI deployment and outcomes.
• A shortage of AI skills and expertise remains a reported adoption challenge.
• Legacy-system and workflow integration can impede implementation.
• Trust, security, privacy, reputational risk, hallucinations and factual-accuracy concerns can limit adoption.
• Some smaller sector, company-size and revenue subgroups have low survey bases.

What to watch

• Whether reported next-12-month productivity gains of 11.7% materialize.
• The persistence of workforce pressure on offshore employees and workers with two to 10 years of experience.
• Whether semiconductor workforce disruption remains greater than in other surveyed sectors.
• Progress in data readiness, AI talent, legacy-system integration and responsible-AI policy adoption.