There is growing concern about artificial intelligence (AI) taking away jobs. The arguments can be categorized into three camps:
- AI will take away close to 50% of the jobs in the next few decades. It is time to take a serious look at solutions like universal basic income.
- AI will take away some jobs, but many more will be created. There is no need to worry; history gives us hope.
- AI will help augment human jobs, rather than replace them. Let us imagine and work toward a better future.
All three scenarios have strong arguments, but the fact that we are debating and talking about it now suggests that, in the end, the more plausible scenario is going to be a mix of all three futures. Tractica’s own analysis suggests that while AI capabilities today are mostly limited, it is a mistake to brush off AI advancement as another hype cycle heading toward an AI winter. Tractica’s recently published report, Artificial Intelligence Market Forecasts, suggests that revenue will reach $36.8 billion by 2025, with almost every industry seeing an impact from AI.
Analyzing the Potential Impact
In this blog post, we try to go a step further by understanding the ratio and mix of how these three scenarios could play out. Compared to some of the existing approaches that take jobs data and analyze the automation potential, Tractica uses a different method by studying the AI use cases themselves. Leveraging our AI forecast model, we analyzed the nearly 200 use cases of AI and estimated their impact on jobs. The result of the analysis suggests that the majority (68%) of use cases for AI have a low probability of impact on jobs; 19% are likely to have a medium impact; and 13% will have a high impact on jobs.
Many of the low-impact use cases will involve AI being used as an augmentation tool, with AI improving the efficiency of the task, enhancing the task, providing new and unique alternatives, or helping to find better processes and procedures to meet end goals. Clinical notes processing for doctors, traffic density predictors for cities, real-time ad performance tools for advertisers, and sports team player selection for coaches are all use cases that will see a low impact from AI in taking away jobs, but rather will see AI used as an augmentation engine. In most cases, the tasks are smaller parts of a bigger workflow, which explains their augmentation potential.
AI will have a medium impact on jobs in approximately 50 use cases. In these cases, while AI could be used as an augmentation tool, it is also likely that the task forms a major portion of the human job description, increasing the likelihood of that job becoming fully automated. This also includes professions that have a relatively fewer number of jobs as part of the overall mix. Examples of medium-impact AI use cases include demand forecasting for supply chains, patient data processing, lead generation for sales and marketing, waste recycling and sorting, medical diagnostics, and music production and generation.
The use cases where AI is likely to have a high impact are fewer, but still significant in number, in the broader context of the 200 or so use cases that were analyzed. We identified 35 use cases in which AI is likely to replace a human worker, such as when the task being automated accounts for the bulk of a human worker’s job description, or workplaces that employ a significant number of people. These specific use cases are likely to cause major societal and political upheaval.
Industries Likely to Be Affected
The bulk of these use cases exist in the business services domain, which includes tasks like customer acquisition and planning, job candidate finders, predictive sales and marketing, travel concierge and booking services, and chatbot-based e-commerce and sales, among others. On the consumer end, use cases like elderly care robots, family robots, and cleaning robots will impact household and care jobs. Healthcare is also likely to see a high impact from AI affecting radiologists, hospital administration, and medical compliance professionals. Doctors themselves are likely to see their roles augmented and to see less demand as AI improves in the areas of diagnosing and recommending treatment.
The advent of driverless taxis, autonomous trucks, and delivery drones will represent a major impact on professional drivers, and is possibly the most visible and politically charged disruption that will occur in the coming years and decade. Stock traders and insurance claims processing jobs are also at high risk, and so are warehouse workers with an increasing number of logistics tasks being automated. Manufacturing jobs are already being automated in countries like China, but with AI, these robots will extend their capabilities and be able to learn skills faster, increasing their usage on the shop floor. Similar to professional drivers, manufacturing is another area that will be highly visible and will have major societal impacts, as a continuation or possible acceleration of the changes that started with industrialization of the economy in the 19th century.
Algorithmic news stories and contract analysis are likely to impact journalism and legal jobs in the future, both being popular professions for highly educated, white-collar workers. Information technology (IT) professionals and software coders are also likely to see an impact with AI being able to perform error checking, but more importantly, with AI able to automatically write the code.
A New Direction for Businesses and Workers
Businesses will become leaner, requiring fewer people at the lower and middle management levels, with many business services like sales, marketing, and human resources (HR) becoming automated. But at the same time, highly skilled professionals like doctors, lawyers, journalists, and IT specialists are likely to see AI impact their jobs. The data shows that AI will not just automate low-level blue-collar jobs like manufacturing, cleaning, or driving. In fact, there are many more AI use cases impacting white-collar jobs than blue-collar jobs, an issue to which policy makers need to pay close attention. It also raises the question of whether enough new jobs can be created, or whether people will be able to transition their skill sets in time to avoid being replaced.
In summary, in the majority of use cases, AI will not impact jobs, and in many instances, it will augment the work humans perform. But, it would be foolish to ignore its disruptive potential, especially in the areas where job descriptions could be fully automated and AI could replace humans. Hopefully, the analysis conducted above contributes to moving this debate forward and finding solutions.