Executive Summary

A recent study by Davenport and Srinivasan challenges the widespread narrative of mass AI-induced layoffs. While 60 percent of surveyed executives report workforce reductions attributed to AI, these decisions are largely based on expectations rather than measurable productivity effects. Only two percent of organizations could trace substantial headcount reductions to concrete AI implementations. The phenomenon is criticized as "AI-washing" – companies justify cost-cutting measures with AI without providing evidence.

People

Topics

  • Artificial Intelligence and Labor Market
  • Corporate Communication and Credibility
  • IT Services Sector
  • Automation Trends Germany vs. USA

Clarus Lead

The central paradox: While major companies like Amazon, IBM, and Klarna justify layoffs with generative AI, measurable productivity gains are lacking. A global management study shows that 60 percent of respondents reduced staff or halted hiring – yet this is based on future forecasts, not realized effects. Relevant for decision-makers: Uncritically linking job cuts to AI harms long-term employee acceptance and jeopardizes productive experimental spaces. In specific sectors such as IT services, however, real effects are evident, while Europe lags significantly behind the USA in automation.

Detailed Summary

The AI-Washing Phenomenon

The research reveals a massive credibility gap: While 39 percent of surveyed organizations reduced staff to a small to moderate extent and 21 percent to a large extent – all in expectation of future AI effects – these decisions are based on speculative scenarios. Particularly critical: Klarna, for example, reduced its workforce by 40 percent between 2022 and 2024 and pointed to AI efficiency gains. In 2025, CEO Sebastian Siemiatkowski acknowledged that the company had transferred too much work to AI, particularly in customer service. This led to public criticism and damages trust in AI initiatives.

Differential Sector Effects

However, the picture is mixed. In the IT services sector – companies such as Tata Consultancy Services, Infosys, and Wipro – real job effects are evident. These firms hired over 100,000 new employees annually at the beginning of the decade; now employment has stagnated as AI-powered programming and automation enable growth without personnel growth. Simultaneously, this has led to an insourcing wave as customers generate AI-produced code more cheaply themselves. In manufacturing, healthcare, and public administration, AI-driven staff reductions are barely measurable.

Regional and Institutional Differences

An OECD study underscores: In Germany, only 12 percent of companies view AI as a near-term replacement technology, while in the USA it is 28 percent. Codetermination rights, strong labor law, and a focus on skills development slow rapid automation in Europe. The evaluation of generative AI itself is also difficult – 44 percent of managers called it the most difficult form of AI to assess, yet 90 percent attested moderate to high overall value contribution.

Scientific Realities

Field experiments show limited, specific effects: Brynjolfsson identified 14 percent productivity gains for junior staff in customer service, with smaller effects for experienced personnel. Microsoft Research documents 10–15 percent gains in software development – under ideal conditions. AI typically automates tasks, not complete jobs; the counterexample of radiologists shows: Despite predictions by Nobel Prize winner Geoffrey Hinton (that AI would surpass radiologists in five years), a decade later there are no lost positions.

Key Statements

  • Credibility Gap: 60 percent of executives justify staff reductions with AI, yet only 2 percent can trace substantial reductions to concrete implementations.

  • Measurability Lacking: Productivity effects are barely quantifiable so far; layoffs are based on future expectations rather than current data.

  • Different Sector Reality: IT service providers actually lose jobs due to AI automation; in manufacturing and healthcare, the effect is minimal.

  • Europe Moving Slower: Only 12 percent of German companies view AI as a near-term replacement technology (USA: 28 percent).

  • Reputational Risk: Uncritical AI-washing damages employee loyalty, increases societal AI skepticism, and jeopardizes productive experimental spaces.

  • Recommended Implementation Approach: Narrow, problem-oriented use cases, incremental staff reduction through natural attrition, process redesign with employee participation, positive communication of relief effects.

Critical Questions

  1. Evidence: Why do 60 percent of surveyed managers report staff reductions due to AI, while only 2 percent can demonstrate measurable implementations? What criteria define "substantial" reductions in the study?

  2. Source Validity: The OECD study cites 12 percent of German companies viewing AI as a near-term replacement technology – how was this sample selected, and are SMEs representatively captured?

  3. Conflicts of Interest: Do companies like Klarna have incentives to justify layoffs on technological grounds rather than admitting to economic pressure or management decisions?

  4. Causality: Can the 10–15 percent productivity gains in field experiments be transferred to real organizations with less ideal conditions? Do the examples (radiologists, software development) also show negative cases?

  5. Feasibility: How realistic is the "narrow and deep" approach with controlled experiments in large, decentralized organizations?

  6. Long-term Impact: If half of Americans view AI use somewhat critically, how does constant AI-washing affect customer trust and consumer behavior?

Bibliography

Primary Source: Amazon, IBM oder Klarna: Waren manche Entlassungen wegen KI zu vorschnell? – FAZ Pro, 05.02.2026, by Holger Schmidt

Secondary References (cited in text):

  • Davenport, Thomas H. & Srinivasan, Laks: Study on AI and Staff Reductions (1,000 surveyed executives, global)
  • OECD Study on AI Automation in Germany vs. USA
  • Brynjolfsson et al.: Field Experiment Productivity Gains (Customer Service, Software Development)
  • Microsoft Research: Productivity Gains Software Development
  • 2025 Survey: American Attitudes Toward AI Use

Verification Status: ✓ 05.02.2026

This text was created with the support of an AI model. Editorial Responsibility: clarus.news | Fact-Check: 05.02.2026