Quick overview: Compared to human labor, AI remains less economical in some industries. There are only 8% of these tasks that are economically feasible for AI automation. Just 0.4% of total worker salaries and benefits in these sectors would be cost-effective to replace with AI.
Developments regarding AI are currently a hot topic. However, this is not commensurate with the costs required to create AI. Many large technology companies do not make a profit from Product AI. This is a big gap, what costs make AI development expensive? Is AI too expensive to replace human jobs? We have done in-depth research on this and found several facts.
Cost Factors That Make AI too expensive
From our findings, the first factor is that the cost of AI becomes expensive when more and more users use AI. Because the large number of users affects the infrastructure costs that must be incurred. So it is not surprising that many users are disappointed with the costs they have to pay to adopt AI modules. Quoted from a statement from Adam Selipsky, the chief executive of Amazon.com’s cloud division, Amazon Web Services, A lot of the customers I’ve talked to are unhappy about the costs that they are seeing for running some of these models.
Apart from that, there are also other factors such as updating AI technology, and creating and training AI, here are the explanations:
Technological Upgraded
AI technology is in a constant state of evolution. Keeping up with these advancements requires significant investment in technological upgrades. These upgrades are essential to enhance AI capabilities, improve efficiency, and stay competitive. However, the costs associated with acquiring cutting-edge hardware, software updates, and integration into existing systems can be substantial, contributing to the overall expense of AI development.
Building and Training AI Products
Developing and training AI models can take years and hundreds of millions of dollars. It involves not only the initial creation of the AI algorithms but also continuous training with large datasets to improve accuracy and functionality. The costs here include not just the technological resources, but also the skilled labor required to design, develop, and train these models. This intensive development process is a significant contributor to the high costs associated with AI products.
Consume Lots of Power
AI systems, especially those involving complex calculations and data processing, consume a substantial amount of power. The energy costs associated with running large data centers and server farms, where AI processing often occurs, are considerable. As AI models become more advanced and require more computational power, this energy consumption – and the associated costs – continue to rise.
Impact AI is too expensive for Business
This not only makes users disappointed with the costs that are too expensive to use AI, but on the other hand, large technology companies are experiencing a dilemma in seeking profits from AI.
Based on reports from Moomoo Technologies Inc., it is reported that Microsoft has still been making losses since its first generative AI product was launched. Quoted from a statement, by Chris Young, Microsoft’s head of corporate strategy, It will take time for companies and consumers to understand how they want to use AI and what they are willing to pay for it.
If it is simulated, an individual pays 10 dollars per month for an AI assistant. In the first few months of the year, the company was losing an average of 20 dollars per month per user. This is because AI models require a lot of power to keep running their programs and require processors that exceed those of regular software or cloud services.
So, Are Humans Still Cheaper Than AI?
With such huge funding for an AI module, does this indicate that humans will still be cheaper than AI? Our findings don’t stop there, the Massachusetts Institute of Technology found a study related to AI replacing humans in a swath of industry.
Study Insights on AI vs. Human Labor in Vision Tasks
The study found that while 36% of U.S. non-agricultural businesses could technically automate at least one worker task with AI computer vision, only 8% of these tasks prove to be economically viable for AI automation. Furthermore, just 0.4% of total worker salaries and benefits in these sectors would be cost-effective to replace with AI.
For large U.S. companies, those with 5000 or more employees, the findings are even more striking. These businesses could cost-effectively automate less than one-tenth of their existing vision-related tasks with AI. This statistic highlights a significant gap in the practicality of AI computer vision integration for most businesses, particularly those of substantial size.
Broader Implications and Future Possibilities
While the study’s findings might offer some reassurance about the current role of human labor, it also opens discussions on other AI applications with potentially lower automation costs. As Gino Gancia from Queen Mary University of London suggests, the landscape of AI integration is varied, and while computer vision might be too costly at present, other areas of AI could offer more economically viable automation options.
Which AI Sector is cheap and can be automated
Quoted from the Bloomberg report, according to Neil Thompson, director of the FutureTech Research Project at the MIT Computer Science and Artificial Intelligence Lab, we show that there will be more automation in retail and health-care, and less in areas like construction, mining, or real estate. These sectors have specific tasks that can be efficiently managed by AI, balancing the cost of implementation and the benefits derived. In retail, AI can revolutionize inventory management, customer service, and personalized shopping experiences. Healthcare, on the other hand, can benefit from AI in areas like patient data analysis, diagnostic procedures, and even routine administrative tasks.
Final Words
The journey of AI development is marked by a complex interplay of high costs and sector-specific challenges. While the potential for AI to revolutionize industries like retail and healthcare is significant, its widespread adoption is hindered by the expenses related to technological upgrades, training, and energy consumption. Currently, in sectors demanding nuanced human skills, AI remains a less economical choice compared to human labor in a swath of industry.
Stay connected for more insights and comprehensive analyses on the ever-changing landscape of AI.
