Why Japanese firms are being so slow to use AI

India’s IT sector continues to thrive despite growing concerns over the impact of artificial intelligence (AI). According to The Economist, the country’s information technology industry is adapting to new technologies while maintaining its core functions. Meanwhile, a separate study conducted in Quebec highlights broader challenges associated with the integration of AI into workplaces, revealing that the benefits of automation are unevenly distributed and often fall short of initial expectations. The Quebec-based research, carried out by a team led by the International Observatory on the Social Impacts of AI and Digital Technology (Obvia), surveyed 4,595 union members representing over 1.4 million workers. The findings suggest that the concept of “augmented work”, where AI enhances human capabilities, does not hold true for all sectors or individuals. Instead, the results reveal a complex landscape where AI either improves productivity, burdens workers further, or leaves them unaffected. Among the key revelations, only 44% of participants reported a genuine increase in productivity after adopting AI tools. Conversely, 26% noted a decline in efficiency, primarily due to the emergence of “AI slop”, low-quality outputs that require additional effort to correct. These outputs include poorly written reports, unverified information, and disorganized communications such as confusing email chains. As a result, some workers spend more time revising or clarifying AI-generated content rather than completing their primary tasks. The study also found that AI does not uniformly reduce workload. While 46% of respondents felt their tasks were eased by AI, 17% experienced an increased burden. In some cases, employees who gained efficiency through AI were assigned more responsibilities, leading to longer hours and higher stress levels. Similarly, workplace stress was reported to fluctuate significantly: 28% of workers said AI reduced their stress, while an equal number claimed it increased theirs. One participant described the rapid evolution of AI tools as “anxiety-provoking and discouraging,” highlighting the emotional toll of constant adaptation. A critical insight from the research is that the effects of AI vary based on factors such as job type, education level, and industry. Professionals and highly educated workers tend to benefit more from AI, often using it autonomously and integrating it into strategic decision-making. In contrast, technical, industrial, and service-sector workers are frequently compelled to adopt AI without clear guidance or training. These workers face heightened exposure to algorithmic monitoring and report fewer tangible advantages from the technology. For instance, 55% of individuals with postgraduate degrees believe they are more productive due to AI, whereas just 22% of high school graduates share a similar view. Another major concern raised by the survey is the lack of organizational oversight and transparency in AI implementation. Only 24% of respondents indicated that their employers had established policies or mechanisms to regulate AI usage. Furthermore, the introduction of AI systems was rarely communicated openly or democratically, with just 12% of participants feeling involved in the process. This absence of structured governance raises questions about how effectively organizations are managing the risks and ethical implications of widespread AI adoption. As the global workforce grapples with these challenges, the Indian IT sector appears to remain resilient, leveraging AI to maintain its competitive edge. Yet, the Quebec study underscores a broader trend: the promise of AI-driven productivity and efficiency is not universally realized. The disparity in outcomes suggests that the success of AI integration depends heavily on context, access, and institutional support. Moving forward, addressing these gaps will be crucial for ensuring that the benefits of AI are equitably shared across all segments of the labor force.