Misconceptions about ChatGPT's Productivity: A Production Management Expert's Perspective

Since the beginning of 2023, with the popularity of Large Language Models (LLMs), an increasing number of people have been making exaggerated claims about productivity improvements using LLMs. Claims of 10x, 20x, and even 100x productivity within a week are rampant, coming from people who do not understand the realities of production management.

There are not a few people who gather many followers by claiming that productivity increases by 10x every day, adding up to 10x, 20x (this is the limit for Kaio-ken), and 30x.

However, production management experts would never make irresponsible statements about how many times productivity or production efficiency will increase due to ChatGPT. Conversely, anyone who claims that productivity will increase by a certain factor without even defining productivity or production efficiency cannot be called a production management expert.Those who make such claims are either amateurs who have never done production management, or malicious information product sellers trying to sell flawed seminar schemes or information products.

Reading the following note will make it easy to understand what a production management expert is. He never talks about airy-fairy stories like productivity increasing by several times due to LLMs. This is an explanation from someone who understands computers and the manufacturing industry.

Below, I will state my opinion on production management in a very simplified manner.

Some reasons why production management experts do not talk about productivity through LLMs

The multifaceted nature of production management:Production management requires integrating many elements, such as logistics, human resources, and equipment optimization, rather than just analyzing simple numbers or data. Although ChatGPT has high capabilities as a language model, it cannot integrate these complex elements to make optimal decisions.

The need for practical experience:Practical experience is extremely important in production management. Models like ChatGPT learn based on massive amounts of data, but they cannot fully understand the unique challenges or subtle nuances of the workplace.

Security issues:Relying on external AI tools for confidential information such as production data and corporate strategy involves security risks. Production management experts prioritize not only improving production efficiency according to departments and processes but also information security and safety.

The existence of specialized tools:Many companies already use specialized software and tools to streamline production management. These tools are specialized for specific industries and production processes and are developed to provide more appropriate solutions than ChatGPT.

The value of human intuition and experience:AI makes decisions based on data, but human intuition and experience are extremely important in production management. Production management experts value that experience and intuition and understand the risks of relying solely on AI.

Differences in productivity between production and non-production departments

Self-proclaimed generative AI productivity improvement consultantsdo not understand the meaning of productivity. To ensure you are not deceived, I will briefly explain the definitions of productivity for production and non-production departments in the manufacturing industry.

1. Definition of productivity in the production department
The production department refers to the departments or lines directly involved in manufacturing products. Its productivity is defined as follows.

Ratio of output to input:Productivity in the production department is often measured by the ratio of how much output (finished goods) can be produced with a certain amount of input (raw materials, labor, energy, etc.).

Relationship with quality:The quality of the manufactured products also affects productivity. For example, if you have high productivity but many defective products, that productivity can actually be said to be low. Terms like defect rate, yield rate, and good product rate are general terms, not technical terms for production management, but if you do not understand this level of common sense, conversation will not be possible.

Reduction of lead time:Productivity is improved by shortening the time from production to completion of a product. In particular, by adopting efficient production methods such as JIT (Just-In-Time) production, improvements in productivity can be expected.

2. Definition of productivity in non-production departments


Non-production departments refer to departments in the manufacturing industry that are not directly involved in manufacturing products, such as human resources, accounting, sales, and marketing. The definition of productivity for non-production departments is as follows.Operational efficiency:
In non-production departments, the ability to perform specific tasks accurately and in a short time is required. Examples include shortening accounting work time and automating sales reports.
Evaluation of contribution:

Measuring how much the activities of non-production departments contribute to corporate profits or brand value as a result leads to the evaluation of productivity.Cost reduction:
Performing work within budget and reducing costs through the elimination of waste are also often evaluated as improvements in productivity for non-production departments.
Productivity in production and non-production departments should be considered on different evaluation axes due to their nature. However, both departments play a role in increasing the overall competitiveness of the company by enhancing their operational efficiency and contribution to the organization.

On production efficiency in non-manufacturing industries

Although there is a clear difference between the manufacturing and non-manufacturing industries, the basic goal of seeking improvements in productivity and efficiency remains the same. In the manufacturing industry, the focus is on bottlenecks in production lines, optimal use of raw materials, and optimization of the workforce.


On the other hand, in non-manufacturing industries, the main focus is on the speed of service provision, quality, and improvement of customer satisfaction. However,

in both industries, basic principles of production management such as appropriate resource management, effective business processes, and continuous improvement efforts are applied. In fact, many non-manufacturing companies and organizations also apply production management methods and tools cultivated in the manufacturing industry to realize operational efficiency and quality improvement. Especially in the modern era, which is an information society where data is at the center of business, the importance of production management concepts and methods is increasing regardless of whether it is the manufacturing or non-manufacturing industry. By incorporating production management concepts across industries and sectors, it becomes possible to enhance organizational competitiveness and support sustainable growth.

Productivity of generative AI in a narrow sense

Please see the following article for the definition of generative AI in a narrow and broad sense.

Generative AI in a narrow senseRegarding this, many people believe that it will significantly improve the productivity of writers, illustrators, and musicians. However, when this is done as professional work rather than as a hobby, it is necessary to consider it from various angles, such as the reliability of the output (especially errors in text created by generative AI), confirmation that the output does not infringe on copyright, whether the quality of the work is maintained, and security. If these factors are not considered, there is a risk that productivity, reliability, and product value will be impaired instead.

If the Tokyo Metropolitan Government changes its way of thinking, it should be able to earn a hundred or even a thousand times more than it does now, which would make a basic income system viable.