Recent data indicates that employment in call centre jobs worldwide has begun to decline, a trend many experts had anticipated with the rise of large language models (LLMs) and other artificial intelligence (AI) technologies. This development is especially significant given the importance of call centres as a source of employment in lower- and middle-income countries (LMICs) and their role as a potential indicator of AI’s broader impact on global labor markets.
A detailed study conducted by Caelan Wilkie-Rogers, an economist at employment research firm Revelio Labs, analysed millions of employment records across dozens of countries to assess changes in call centre employment since late 2023. The findings reveal a 5 percent global decrease in call centre jobs from their peak around that time. More strikingly, this decline is evident not only in high-income countries—where call centre roles have been diminishing for some years due to offshoring—but also in many LMICs, such as the Philippines, which has seen a 10 percent reduction, and Kenya, with a 6 percent drop over the past 18 months.
The timing of these employment shifts suggests that AI may now be playing a material role in accelerating job losses. Call centre positions are particularly vulnerable to AI disruption because they rely heavily on routine, well-defined tasks that language models efficiently handle. Moreover, the outsourced nature of many call centre roles weakens employer-employee bonds, making these jobs more susceptible to automation.
However, the study underscores that the reduction in call centre employment primarily reflects a slowdown or halt in new hiring rather than widespread layoffs. In fact, layoffs have decreased, albeit less sharply than hiring rates, suggesting employers are downsizing quietly through attrition.
Interestingly, the data also points to resilience and growth in higher-paying, white-collar roles within LMICs, even those theoretically exposed to AI. Employment in roles such as client relations and other advanced customer service functions has increased. Yet, it appears that workers shifting into these higher-value roles are generally not the same individuals leaving entry-level call centre positions. Instead, those exiting call centre jobs often move into lower- or similar-paying occupations.
These patterns highlight several broader themes: job displacement is influenced not only by how closely tasks overlap with AI capabilities but also by the socio-economic status and security of the occupations. Lower-paid, outsourced jobs are disproportionately affected. Furthermore, the most visible impact of AI on vulnerable roles is through reduced hiring rather than mass unemployment. Lastly, signs of AI cutting hiring in exposed jobs coexist with positive net effects, possibly fueled by increased demand and AI’s capacity to lower skill thresholds.
This trend raises concerns for economies that rely heavily on call centre employment. For example, in the Philippines, the outsourcing sector contributed approximately 7.4 percent of GDP in 2023. Experts caution that automation of entry-level office jobs could impede traditional pathways for upward mobility, a phenomenon likened to “premature deindustrialization.” At the same time, the growth of higher-skilled roles in countries such as India offers some cause for optimism, though it remains unclear whether these positions will serve as alternative entry points for displaced workers.
Beyond call centres, recent advances in AI demonstrate the technology’s expanding reach into complex, specialist fields. For instance, OpenAI recently announced that a new, unreleased model has solved numerous longstanding mathematical problems, provoking a mix of amazement and concern among professional mathematicians. This breakthrough exemplifies AI’s potential to transform knowledge-intensive work and raises questions about the future role of human expertise and skill development in such domains.
As AI continues to reshape various labor markets, ongoing research and observation will be crucial to understanding which jobs are most vulnerable, how workers can transition, and what policies may be required to mitigate the socio-economic impacts of automation.