Technology is raising productivity for some workers already inside the digital economy, while weak local-language data, infrastructure gaps and limited adoption have so far constrained its impact on employment.
In May 2026, Abbay Television advertised an entry-level position in Addis Ababa with a job title that would have been unusual only a few years earlier: AI Content Creator Intern. The vacancy asked applicants to generate prompts from scripts and ideas, create and refine AI-generated material for broadcast and digital platforms, and carry out basic editing and quality checks.
Two months later, Ethiopian technology company AQEM AI promoted an advertising tool that it said could generate short social-media videos with Habesha-looking virtual presenters speaking Amharic. The company's marketing material said a business could describe a product, select a presenter and scene, and produce a video without arranging an actor, studio or conventional editing workflow.
Neither example establishes that jobs have been lost. There is no national dataset showing that Ethiopian advertising agencies, media companies or technology firms have reduced employment because of generative AI. The examples instead show two changes occurring at the same time: software is beginning to perform parts of existing production processes, while employers are also creating roles for people who can operate AI tools.
An IT executive who has run an Ethiopian technology firm for more than a decade and asked not to be identified said the current level of adoption is too limited to describe AI as an economy-wide employment shock.
"It cannot be said that AI has shrunk the job market," he said. "To begin with, we are not a country that is part of the AI revolution. Fundamentally, AI is greatly supporting the people who are already within the ecosystem. However, someone who is outside the system is not benefiting."
The employment question is critical in Ethiopia because the country was already struggling to create enough jobs before the recent expansion of generative AI. In June 2026, the World Bank placed job creation at the centre of its new country strategy for Ethiopia, saying more than two million Ethiopians enter the labour market every year.
That demographic pressure makes the Ethiopian debate different from the discussion in many advanced economies. In countries with large formal service sectors, much of the concern has centred on whether AI will replace existing salaried workers. In Ethiopia, an additional question is whether technology could change the number and type of new positions created as businesses expand.
A company does not have to dismiss workers for technology to affect employment. If a growing department can handle additional work with the same staff using AI-assisted systems, future recruitment may be lower than it otherwise would have been. At the same time, lower costs can allow firms to expand, enter new markets and hire in other functions. Ethiopia does not yet have firm-level data showing which of those effects is dominant.
World Bank sees lower automation exposure in developing economies.
The World Bank's World Development Report 2026 estimates that 4.5 percent of existing jobs in low- and middle-income economies are at risk of automation by generative AI, compared with 14.2 percent in high-income countries. It estimates that 16.2 percent of jobs in developing economies could receive a meaningful productivity boost from AI.
The International Labour Organization has reached a similar broad conclusion using a different methodology. Its research finds that exposure to generative AI is lower in low-income economies than in high-income countries and that transformation of jobs is more likely than full replacement. Clerical occupations are among the most exposed, while many manual and agricultural tasks remain less susceptible to current generative systems.
Ethiopia's employment structure helps explain the lower immediate exposure. Agriculture, informal trade, construction, transport and other occupations requiring physical presence account for a large share of work. The country has a smaller share of the highly digitised office occupations that current generative systems can most easily automate.
The World Bank also identifies a second side to the gap: developing economies may capture less of AI's productivity benefit if workers and firms lack reliable power, connectivity, computing infrastructure and skills. Its 2026 report frames access to those foundations as a major factor in determining how widely AI gains spread.
For some workers, AI is already a productivity tool
The anonymous IT executive said he sees the clearest effect inside software development, where AI is shortening tasks rather than eliminating the role of the developer.
"For a website developer, giving me AI is like giving me a machine that understands my needs," he said. "It saves me time and effort."
Beakal Gizaw, a computer-science lecturer at Addis Ababa University, made a similar distinction in an interview for this story. He said firms can obtain more value by combining workers with AI than by treating the technology mainly as a substitute for labour.
"The modern approach is not to replace people with AI," Beakal said. "It is to give AI to employees as an assistant. The strongest results come from organising people and technology together, rather than replacing people with technology."
Beakal also pointed to institutional knowledge as a constraint on rapid replacement. Employees carry knowledge about clients, suppliers, past decisions and internal processes that may not be captured in a company's digital systems. He said large-scale displacement can also affect morale and organisational performance.
Global research gives mixed signals on headcount.
Evidence from larger international companies does not show a single employment pattern. Gartner surveyed 350 executives at companies with at least one billion US dollars in annual revenue that were piloting or deploying AI agents, intelligent automation or other autonomous technologies. About 80 percent reported workforce reductions, but Gartner found that staff reductions were nearly as common among firms reporting high returns as among firms reporting modest or negative returns.
PwC's 2026 Global AI Jobs Barometer points in a different direction at companies with high exposure to AI. The study, based on more than one billion job advertisements across six continents and company financial data, reported that headcount at the most AI-exposed companies had increased 52 percent from a 2018 baseline, compared with 36 percent at the least AI-exposed companies. Jobs requiring specific AI skills were growing faster than the overall jobs market, while advertised roles requiring AI skills carried an average wage premium of 62 percent.
The studies measure different groups and different forms of AI adoption, so their findings are not directly comparable. Together, they show that the use of AI does not automatically translate into either lower or higher employment. Outcomes vary with business expansion, investment, workflow changes and the type of work being automated or augmented.
Entry-level work is one area to watch.
PwC's 2026 analysis also found a change in the skills requested for junior positions. In its U.S. data, AI-exposed entry-level roles were seven times more likely than less-exposed entry-level roles to request skills traditionally associated with more experienced workers, including judgement and leadership. The finding does not measure Ethiopia, but it illustrates how automation of routine tasks can alter the content of junior jobs.
Meron Tesfahun, a 24-year-old management graduate, had applied for more than 70 jobs over two years. Her experience predates any evidence that AI is affecting recruitment in Ethiopia and cannot be attributed to the technology. It illustrates the difficulty many graduates already face in reaching the first rung of formal employment.
If Ethiopian employers begin using AI to automate routine research, drafting, data entry or basic production work, the effect on entry-level recruitment would need to be measured separately from layoffs among existing employees. No national data currently track that distinction.
Experts say Ethiopia still lacks evidence of an employment effect.
Bahiru Abas, an AI trainer and specialist who works with companies implementing AI, said Ethiopia remains at an early stage of adoption and that available information does not support claims of either widespread job losses or widespread job creation.
"In Ethiopia, the state of AI has not advanced beyond a novelty stage," Bahiru said. "There is insufficient data to say whether employment levels have increased or decreased."
The anonymous IT executive offered a similar assessment from the private sector.
"As for whether it has affected the overall job market, my answer is that it hasn't touched it at all," he said. "There is still a long way to go before we can say the job landscape has been affected or not."
His statement is an assessment based on his experience in the technology industry rather than a national employment survey. Ethiopia does not currently publish regular firm-level statistics showing AI adoption by sector, changes in recruitment linked to AI, or the number of positions eliminated or created because of the technology.
Local-language data remain a constraint.
Language is one factor limiting the spread of generative AI across Ethiopian workplaces. GSMA research on AI use cases in Ethiopia says only about 0.02 percent of internet content is in African languages, compared with about 53 percent in English. Ethiopia has more than 80 languages, and digital material for many of them remains limited.
The IT executive pointed to translation as an example of the difference he sees between international AI adoption and the Ethiopian market. He said AI has altered translation work internationally, while local-language translation remains more dependent on human workers in Ethiopia.
"In Ethiopia, translation work still exists. Translators are still active," he said. "The algorithm feeding for both Amharic and Oromo is very weak. Moreover, there are no other stories or robust data sets to draw from."
Research by GSMA similarly identifies shortages of local-language training data as a constraint on AI systems in Ethiopia. The problem affects applications including chatbots, voice assistants, financial services, education tools and agricultural advisory systems, where accuracy depends on the system's ability to understand the language and context of users.
Connectivity and electricity shape access.
The broader digital infrastructure also limits how widely AI tools can be used. The 2025 Network Readiness Index ranked Ethiopia 119th out of 127 economies. It placed the country 121st on its people pillar, 124th on governance and 119th on future technologies.
Telecommunications coverage has expanded. Ethio telecom reported that its 4G population coverage reached 82.23 percent in the 2025/26 financial year, while Safaricom Ethiopia reported 59 percent 4G population coverage by the end of March 2026. Measures of actual internet use remain lower: DataReportal estimated 29.5 million internet users in Ethiopia at the end of 2025, or 21.7 percent of the population.
Electricity access is another constraint. Ethiopia's National Energy Compact, citing the 2025 Ethiopian Energy Access Survey, says 65 percent of households had access to at least one source of electricity as of January 2025, while 44 percent of the population had basic Tier 1-or-above electricity service.
The gap means access to AI is not uniform across workers or firms. People with reliable connectivity, devices and paid digital services can use AI tools more consistently than workers and small businesses without those inputs.
Data sovereignty is also part of the debate.
The IT executive also raised concerns about where Ethiopian data are stored and where computing capacity is controlled.
"The overall issue of sovereignty is concerning," he said. "If the narrative is not handled carefully now, it will be difficult later. We need to build our own data center."
Ethiopia already has domestic data-centre infrastructure, and the government is planning additional sovereign computing capacity. Digital Ethiopia 2030 calls for regional data centres, national AI computing infrastructure and integration with sovereign cloud systems. In June 2026, government officials also said a national data cloud was under development at the Ethiopian Artificial Intelligence Institute.
The National Artificial Intelligence Policy, approved in 2024, also addresses data, infrastructure, human capacity, research, governance and ethics. The policy recommends investment in a national computing grid and data warehouse while seeking to attract international cloud and infrastructure providers.
Government response focuses on skills and domestic capability.
Ethiopia established the Ethiopian Artificial Intelligence Institute in 2020 and approved its National Artificial Intelligence Policy in June 2024. In February 2026, the Institute, Addis Ababa University and the United Nations Development Programme inaugurated the AI University Innovation Pod in Addis Ababa, with high-performance computing systems, robotics equipment, laboratories and research spaces.
The government has also used mass digital training as part of its response. The 5 Million Ethiopian Coders Initiative, launched in July 2024, recorded 5,005,146 enrolments by June 11, 2026 across programming, data analysis, Android development and artificial-intelligence fundamentals, according to the Government Communication Service and Ethiopian News Agency.
The figure measures enrolments, not employment. Public announcements on the milestone did not provide a national figure for how many participants obtained technology jobs, freelance income or business revenue as a result of the programme. The government has said participants receive certification and has increased its enrolment ambition beyond the original five-million target.
Experts also raise distributional concerns.
Natenael Hailu, a digital-economy expert interviewed for the story, said the public discussion around AI can overstate the technology's current capabilities while paying less attention to who has access to it.
"The excitement around AI has gone too far," Natenael said. "Many organisations are portraying its capabilities as far greater than they really are."
He said the distribution of access could widen differences between workers and firms. Larger businesses are better positioned to pay for computing, software and specialised staff, while workers with stronger digital skills can use AI to increase their output. Those without connectivity or relevant skills may see less immediate benefit.
The ILO has also found that exposure to generative AI is uneven across occupations and gender globally, with clerical work among the most exposed categories. Ethiopia does not yet have sufficiently detailed national evidence to quantify how those patterns are playing out locally.
Employment impact remains unresolved.
As of August 2026, the available evidence does not show that AI has reduced Ethiopia's overall job market. Local experts interviewed for this story say adoption remains limited, and national data do not isolate AI-related changes in employment.
There are, however, observable changes at the task and skills level. Ethiopian businesses are being offered software that can automate parts of advertising and content production. Employers have begun advertising AI-specific roles. Software developers report using AI to shorten work, while international research shows that employers in more AI-exposed markets are changing the skills they request from junior workers.
The direction of Ethiopia's employment effect will depend on how widely firms adopt the technology and whether productivity gains lead mainly to lower labour requirements, business expansion, new products or a combination of those outcomes. At present, the data are not sufficient to determine the balance.
For a labour market that must absorb more than two million new workers every year, the impact of AI cannot be measured by layoffs alone. The larger question is whether firms adopting the technology will expand their workforce, hire more selectively, or require fewer people as they grow. In Ethiopia, there is not yet enough evidence to answer that conclusively.
