AI is already rapidly shaping our everyday lives. For example, many of us rely on AI for translations. But what are the wider implications for work, employment and the economy? In this exclusive essay for CONVOCO! – Carl B. Frey, Professor of AI and Work and Director of the Future of Work Program at Oxford University – summarizes key findings from his recent research on:
AI and the Need for Translations and Foreign Language Skills
Rapid advancements in artificial intelligence (AI) are significantly reshaping many professions, with machine translation (MT) technologies emerging as a prominent example. While AI offers increased accessibility and the potential to overcome language barriers, it simultaneously raises critical questions about the future employment prospects for human translators and the overall economic demand for foreign language skills.
The evolution of translation technology has witnessed several notable milestones. The journey began with early experiments such as the IBM701 machine in 1954, marking the first significant attempt to automate translation. During the Cold War era, machine translation garnered considerable attention due to strategic interests in translating Russian scientific documents into English. Despite initial optimism, early systems faced considerable limitations, highlighted in the influential 1966 ALPAC report, which questioned the feasibility of MT, leading to reduced research funding and a temporary decline in interest.
Machine translation experienced a renaissance in the late 20th century, particularly driven by advancements in computing power and internet accessibility. In 1997, Babel Fish became one of the earliest widely accessible online translation services, significantly impacting public perception of MT’s practicality. The subsequent launch of Google Translate in 2006 introduced statistical machine translation, utilizing vast amounts of bilingual text data to improve translation accuracy dramatically.
Google’s introduction of Translate as a mobile application in 2010, integrated seamlessly into browsers like Chrome, significantly accelerated its widespread adoption. This technological advancement reshaped how societies interact across language boundaries, fostering unprecedented global connectivity and information exchange. Between 2012 and 2021, these developments propelled the U.S. translation industry’s growth from $33.5 billion to $37 billion. Yet, despite this economic expansion, significant concerns emerged among human translators. A 2024 survey indicated that over 75% of translators feared generative AI would negatively impact their incomes, reflecting broader anxieties about the future relevance of their skills.
In our study, we explored how AI affects translator employment by examining the experiences of nearly 700 cities across the United States. By connecting local translator employment data with regional interest in Google Translate, we uncovered a clear and compelling pattern. Each small increase in the use of machine translation was associated with a noticeable reduction in jobs for translators and interpreters. In practical terms, this meant approximately 28,000 fewer translator jobs emerged between 2010 and 2023 than might have been expected without widespread AI adoption. These findings highlight a vivid picture of how swiftly technology can alter employment landscapes.
The impacts extend far beyond translators alone. Historically, fluency in foreign languages has been prized across fields such as international business, healthcare, customer service, and education. Yet, our analysis of millions of job advertisements across different regions shows a marked slowdown in demand for bilingual skills in areas heavily adopting machine translation. For instance, regions with high usage of machine translation saw significantly fewer job postings requiring proficiency in Spanish, Chinese, German, French, and Japanese. Interestingly, sectors like IT and engineering experienced less severe declines, suggesting that advanced technical roles still highly value language proficiency for nuanced, complex interactions.
The broader economic implications of our findings are significant, particularly concerning global trade. Historically, linguistic barriers have posed substantial obstacles to international commerce, incurring costs comparable to those of formal trade restrictions such as tariffs. By lowering these barriers, improved MT has the potential to stimulate international trade, especially in services. This advancement is particularly promising for developing countries, which traditionally relied on manufacturing exports but now face limited growth opportunities due to premature deindustrialization driven by automation and robotics.
AI-enhanced translation capabilities can empower small businesses and professionals in non-English-speaking regions, such as Nairobi or Dhaka, to participate more effectively in global markets. By leveraging MT tools, these businesses can bid on international projects, engage in global supply chains, and offer professional services worldwide, facilitating greater economic inclusion and diversification into higher-value service sectors.
Here you can find the full research paper by Carl B. Frey and Pedro Llanos-Paredes