Author: Maggie Smith
Generative AI is a subset of artificial intelligence technologies that creates new, original content based on patterns it learned from a vast amount of training data. Perhaps the most well-known example is ChatGPT, which gained a remarkable 100 million users in just 60 days. With the advent of this new technology, many have questioned how AI may impact the labor market. While this is still a new and ever evolving topic, here are the top five things we have learned about generative AI and the workforce after our first LEAD article on ChatGPT.
AI may impact jobs that were previously thought to be safe from automation.
AI was developed to replicate and extend human cognitive functions through advanced technologies like machine learning and natural language processing. Unlike traditional automation, which is designed to perform repetitive tasks, generative AI excels in creating new content, processing large amounts of data, and facilitating decision-making. Its capacity to learn and adapt makes AI capable of reshaping and automating a broader range of tasks than traditional automation. Thus, AI has the potential to influence jobs traditionally considered immune to automation, including roles requiring creativity or complex cognitive skills.
Recent studies leveraging O*NET data have attempted to predict which occupations are most susceptible to AI disruption. This research evaluates the likelihood of AI impacting jobs by comparing the tasks and skills required in each occupation with what AI can perform. A significant finding across these studies is the potential for AI to influence white-collar or professional jobs, while the least exposed occupations tend to be physical and/or outdoor occupations (see Figure 1). This suggests that occupations that typically require higher education and offer higher pay are prone to technological disruption by AI.
However, it is unknown to what degree these occupations may be impacted. There is still a lot to learn, and many factors must align. For example, more work needs to be done to fully understand the reliability of the technology for industry-specific applications. AI adoption will also be influenced by the workforce skills needed to utilize the technology and the willingness of employees to integrate AI tools into their workflow. Also, broader societal acceptance and ethical considerations will play a role. These are just a few examples of factors that will likely contribute to the ecosystem necessary for AI to have a transformative effect on the workforce.
AI may be used to augment and improve work.
AI may increase job satisfaction by automating mundane tasks and enabling focus on meaningful work. Research like Pizzinelli et al, 2023 incorporates O*NET metrics to identify occupations AI might enhance rather than automate, focusing on roles requiring in-person interaction, critical decision-making, and specialized expertise. This approach identified professions such as lawyers, surgeons, and judges as highly likely to be complemented by AI, underscoring its potential to support rather than replace human expertise.
Further, AI may be able to complement other jobs, such as through customer service chatbots that handle routine inquiries, allowing human agents to focus on more complex customer needs. In retail, AI-powered inventory management systems can predict stock levels, freeing up staff for customer-facing roles. Similarly, in manufacturing, AI may be used to monitor equipment performance to predict maintenance needs, reducing downtime and allowing workers to focus on production rather than reactive maintenance tasks.
Thus, the integration of AI presents an opportunity to enhance the quality and efficiency of work across a multitude of sectors. By automating routine and mundane tasks, AI can potentially allow workers to focus on the core aspects of their roles that require human insight, creativity, and specialized skills.
AI may have positive impacts on productivity and democratization of skills.
Several studies have suggested AI may be beneficial for worker productivity across tasks like business writing, programming, customer support, and consulting. These studies compared groups with and without AI assistance on the number of tasks completed, time to complete tasks, and in some cases the quality of output. Notably, the most significant benefits in productivity were observed among less-experienced, lower-skilled workers, although AI tended to improve performance for all.
Thus, workers who learn to use AI tools may be at an advantage, particularly less-experienced workers. Although training is advisable to effectively leverage the technology, both in formulating effective prompts and grasping its limitations, this technology is widely accessible, given that programming skills are not needed. Indeed, there has been discussion about the potential for AI to democratize access to knowledge and skills.
AI still has significant limitations.
Generative AI, and commercial AI products, have seen significant improvements in capability and usability in the past year. But while powerful and full of potential, they still possess significant limitations. These limitations include a propensity for bias in the outputs, reflecting biases present in training data. Generative AI systems can also produce unpredictable or nonsensical results due to their reliance on patterns in data rather than true comprehension. Additionally, generative AI is susceptible to hallucinations, or instances where artificial intelligence systems generate false or misleading information, often because of misinterpreting their training data or attempting to fill gaps in their knowledge.
Recognizing limitations of generative AI is important, but it should not deter leveraging its capabilities. Understanding these boundaries allows for more informed and strategic application of AI, ensuring its benefits are maximized while mitigating potential risks.
And finally, more research is needed.
Our understanding of how AI may impact the labor market continues to evolve and requires ongoing research to fully understand the implications. Future studies should not only assess AI's technological capabilities but also consider its ethical, economic, and social impacts, ensuring that its deployment will be beneficial, not harmful.
At LEAD, we're dedicated to staying up to date on the latest AI research and trends in the labor market, ensuring we share our insights to maintain a well-informed community across North Carolina. While ChatGPT and other commercial AI products are widely available, we have not seen the fast and widespread changes in the labor market as some had initially feared. There is potential for acceleration, but for now it seems more likely that the impacts, positive or negative, may take time to transpire.