This post was originally published on this site.
As artificial intelligence (AI) becomes more widespread, fundamental transformations are occurring in processes and the skill sets of the workforce across all affected domains. In particular, concerns in labor markets are growing due to the expectation that the expansion of this technology may lead to job contractions. AI is not only taking over roles subject to automation but also engaging in non-routine cognitive tasks such as abstract reasoning, prediction and judgment. As a result, this concern extends from whether existing employees can retain their positions to whether new graduates can meet the expected skills in their job searches. Naturally, these trends vary depending on the economic and labor market dynamics of different countries. However, this uncertainty and apprehension are pushing many businesses, even those not directly related to AI, to prioritize hiring individuals with AI-related skills.
When these technologies continue to progress with their current fluidity, the concerns being felt are highly justified. This is because AI rapidly takes over roles susceptible to automation across all professions and job positions. Consequently, business processes and job roles are being redefined, workforce displacement within organizations is increasing, and ultimately, the number of employees losing their positions is rising. As traditional job roles decline, the skill expectations for new job positions are continuously increasing. For example, in China, where significant investments are being made to lead in this field, many companies now require a Ph.D. degree for hiring, rather than simply valuing experience. As this trend strengthens, workers with low and medium skill levels are being displaced by advancing automation. Since they lack the high-level skills needed to match emerging job positions, they are pushed out of the labor market. If fewer but highly skilled workers become sufficient for new business processes, the human-machine balance in the labor market will shift, to the detriment of human workers.
On the other hand, recent studies indicate that the potential for these technologies to complement human labor presents a highly promising alternative path. In this context, generative AI technologies have been shown to lead to significant improvements in employee performance, particularly by enhancing the productivity of lower-performing workers, ultimately resulting in notable gains in overall efficiency. For instance, a study conducted using GitHub Copilot, a generative AI-based programming assistant, demonstrated that the experimental group with access to Copilot completed programming tasks approximately 56% faster than the control group without access. Subsequent studies have made it even more apparent that AI’s positive impact is not distributed homogeneously across all workers.
In this context, one of the first studies examining the effects of generative AI tools analyzed data from 5,179 customer support representatives to investigate the impact of a generative AI-based chat assistant on employee performance. The study found that the AI-powered chat assistant increased productivity by an average of 14%. More interesting findings emerged when analyzing how this improvement varied based on employees’ skill levels. The impact differed across skill levels, with newly hired and lower-skilled employees experiencing a productivity increase of 34%, significantly above the average. In contrast, the effect on experienced and highly skilled employees remained minimal. In other words, the positive impact of generative AI on employee productivity diminishes as experience and skill levels increase.
Another study examining the impact of ChatGPT on writing tasks performed by university-educated employees, such as press releases, short reports and analysis plans, found that ChatGPT significantly increased writing speed by 40% and output quality by 18%. However, the study also showed that the most significant benefits were observed among lower-skilled writers. A similar survey of management consultants found that employees using this technology not only completed 12% more tasks but also reduced task completion times by 25% while improving output quality by more than 40%. The findings further revealed that lower-skilled employees benefited the most from these improvements, leading to a narrowing of the productivity distribution.
More efficiency
In short, AI technologies function as a complementary support mechanism, enhancing employee efficiency and helping workers perform their tasks more effectively by narrowing the productivity distribution. In other words, AI significantly improves the performance of lower-performing employees in terms of both output quality and production time, thereby increasing overall productivity and efficiency. This finding challenges the widely accepted notion that new technologies primarily complement highly skilled workers. Instead, as demonstrated by recent studies, AI technologies tend to complement lower-skilled employees. Another important insight from these studies is that new employees using generative AI tools can reach a previously time-intensive proficiency level in a much shorter period. In other words, generative AI tools such as ChatGPT can significantly reduce the time required for beginners and apprentices to develop their skills.
The studies mentioned above on the impact of generative AI on productivity and efficiency have primarily been conducted in the private sector. To address this gap, new research is now being carried out to examine its effects in the public sector. For example, a recent study investigated the impact of generative AI technologies on document comprehension and data analysis tasks performed by employees at the Central Bank of Ireland. The findings revealed that for document comprehension tasks, AI improved quality by 17% and reduced task completion time by 34%. However, no positive contribution was observed in data analysis tasks. Consistent with previous studies, lower-performing employees experienced the greatest benefits in document comprehension. These findings suggest that while generative AI can significantly enhance quality and productivity in the public sector —despite its structural and procedural differences from the private sector — the extent of its positive impact depends on the nature of the task.
More detailed studies on the impact of AI on employee performance, particularly those considering the varying characteristics of tasks (such as simple vs. complex or explicit vs. ambiguous), provide valuable insights for mapping this effect. In this context, a recent study conducted with a representative sample of the working population in the United Kingdom examined the impact of generative AI (ChatGPT) on employee performance based on task characteristics. The study found that generative AI improved employee performance in terms of both time savings and quality, regardless of gender, education level or professional background. However, the effect was significantly more pronounced in complex but explicit tasks. In contrast, AI had little impact on simple tasks, while in complex and ambiguous tasks, employee skills remained the primary determinant of performance.
The impact of generative AI applications on performance and quality is also evident in scientific research and article writing. Researchers quickly recognized the opportunities offered by generative AI tools such as ChatGPT and began incorporating them into academic writing. In fact, articles listing AI as a co-author have already started to appear. This development has prompted scientific journals to debate their policies on the matter. The central issue in these discussions has been the capacity to assume responsibility, leading to a consensus that AI cannot bear responsibility in this context. Initially, most journals, including Science, took a strict stance, not only prohibiting AI from being listed as an author but also rejecting the use of AI-generated content in scientific publications. However, over time, many academic journals have gradually relaxed their position on this issue.
At this stage, while it is generally accepted that AI cannot be considered an author in a scientific article, there is also broad recognition that these technologies can make significant contributions to scientific research and article preparation. As a result, it is now widely acknowledged that such contributions should be explicitly stated in publications, as they are expected to enhance the overall quality of academic work. In recent publications, AI contributions have been increasingly disclosed. With the introduction of DeepSeek’s new platform, these contributions are expected to grow even further. As in other fields, younger researchers and those with language barriers tend to benefit the most from AI-assisted writing. Similar benefits can be realized in education systems, particularly in helping students close achievement gaps through personalized learning and supporting teachers in professional development programs.
In summary, AI technologies have been shown to contribute positively to employee productivity and efficiency. However, the extent of this contribution varies depending on the nature of the tasks. AI’s impact is more pronounced in non-routine cognitive tasks and complex but explicit assignments. Additionally, new and lower-performing employees benefit the most from these improvements. With the assistance of AI, new hires can reach the expected competency levels much faster. The findings from these studies suggest that, rather than negatively impacting employment, the widespread adoption of AI technologies can enhance the performance of lower- and middle-skilled workers in the workplace. By prioritizing AI as a complement to human labor rather than a driver of automation, businesses can increase productivity without reducing employment, thereby fostering a more equitable distribution of economic benefits and strengthening the middle class. However, realizing these benefits will require strong policy support for AI as a complement to human labor, as well as its integration into business management strategies.