Archive · 19 Aug 2026
Goldman Sachs Warns AI May Reshape Jobs, Especially Entry-Level Work
Goldman Sachs studied how artificial intelligence may change jobs around the world.
It found that some industries using AI have had slower job growth since 2022.
Call centers, software publishing, consulting and advertising are among the industries facing pressure.
Young people starting their careers may be affected most because AI can do some tasks once given to junior workers.
About 300 million jobs worldwide include tasks that AI might perform.
This does not mean 300 million jobs will vanish.
AI can also help workers and create new jobs, including jobs building data centers and power systems.
Some future jobs may involve managing AI, while others may appear because people have more income and demand new services.
The size of the disruption will depend partly on how quickly businesses adopt the technology.
Goldman Sachs says AI-exposed industries have seen weaker job-opening growth since the second half of 2022, especially in Germany, Australia and the US.
Employment in call centers is estimated to be 39% below trend in the US, 33% below trend in Canada and 27% below trend in Germany.
Entry-level workers face the strongest AI-related headwinds, with a 10% occupational exposure linked to a larger employment-growth drag for new entrants.
About 300 million jobs globally are exposed to AI automation, but exposure does not necessarily mean entire occupations will disappear.
AI infrastructure is also creating jobs, including 216,000 additional construction roles linked to data-center development since 2022.
- Who
- Goldman Sachs Research, led in the report by Joseph Briggs and analyst Evan Tylenda, assessed the effects of artificial intelligence on workers and industries.
- What
- The research found early signs of weaker hiring and employment growth in some AI-exposed industries, particularly for entry-level workers, while also identifying potential new jobs.
- When
- The analysis focuses largely on changes since the second half of 2022; Goldman says 2026 could be an important year for the US labor market.
- Where
- The findings cover major developed economies, including the US, Germany, Australia, Canada, France, the Netherlands, the UK, Italy, Japan and New Zealand.
- Why
- Businesses are increasingly adopting AI to perform or assist with human tasks, potentially changing hiring needs, worker skills and the types of jobs created.
AI Displacement Risks
AI Job-Creation Potential
Overall employment effect
AI Displacement Risks
AI may reduce hiring and displace workers, particularly in call centers, consulting, graphic design and other knowledge-based occupations.
AI Job-Creation Potential
AI may augment workers, raise productivity and create new specialist roles or jobs driven by higher incomes and demand.
Entry-level opportunities
AI Displacement Risks
Companies may need fewer junior employees if AI performs routine writing, research, analysis, customer-support, coding and design tasks previously used to train new workers.
AI Job-Creation Potential
New opportunities may emerge for workers who develop AI-related skills and for specialists who manage, implement or work alongside AI systems.
Economic transition
AI Displacement Risks
Workers leaving declining knowledge occupations may not immediately have the skills required for expanding technical roles such as engineering, construction or electrical work.
AI Job-Creation Potential
AI-related infrastructure is already increasing demand for construction, electrical, HVAC and other technical workers, offering new employment pathways.
- Jobs exposed globally
- Around 300 million jobs are potentially exposed to AI automation.
- US work hours
- AI could potentially automate tasks accounting for 25% of all work hours in the US.
- Call-center employment
- Employment is estimated to be 39% below trend in the US, 33% below trend in Canada and 27% below trend in Germany.
- Entry-level impact
- A 10% occupational exposure to AI was associated with an employment-growth drag exceeding 0.6 percentage point in Australia and 0.2 percentage point in the US for entry-level workers.
- Potential displacement
- Goldman estimates that 6%-7% of workers could eventually be displaced in its base case.
- Data-center construction
- Construction jobs exposed to data-center development have increased by 216,000 since 2022.
- US power-sector demand
- Roughly 500,000 net new jobs may be needed in the US by 2030 to meet growing power demand.