The manufacturing industry has embraced modern technologies such as big data, machine learning, and artificial intelligence. This paper examines AI and machine learning developments in the manufacturing industry, comparing current practices and data-driven projects. It aims better to understand these technologies and their potential benefits and challenges. The research identifies opportunities for innovative business solutions and explores industry practices and research results. The paper focuses on implementation rather than technical aspects, aiming to enhance knowledge in this area.

AI and ML are bringing a major revolution to the manufacturing industry. According to a study, AI is expected to impact the industry by 2023 significantly. ML can create efficiencies and analyze complex data sets quickly (As shown in Figure 1).

Figure 1

Introduction to AI

AI enables computer systems to perform tasks that require human intelligence. ML is closely related to digital twinning, which improves performance and future designs. This reduces time, cost, and resource needs in the manufacturing process.

1.1. Background

The Fourth Industrial Revolution is characterized by new technologies that merge the physical, digital, and biological worlds. Artificial Intelligence (AI) is a driving force in this revolution, with the potential to revolutionize industries. Well-designed AI systems can learn and improve over time, leading to more innovative products and services. The manufacturing industry is particularly interested in AI, which contributes significantly to the economy. This paper aims to explore the adoption and applications of AI in manufacturing, providing insights into its impacts and challenges [1, 5] (As shown in Figure 2).

Figure 2

AI in Manufacturing Industry

1.2. Problem Statement

AI and ML have the potential to revolutionize manufacturing by enhancing efficiency, productivity, and innovation. However, leveraging these methods for optimization in product design, automation, maintenance, and supply chain management is challenging due to their novelty and need for more extensive knowledge. Manufacturing faces unique challenges in using AI and ML methods due to the diversity and complexity of the field [2]. This study aims to analyze the use of AI and ML in manufacturing, identify the benefits and obstacles, and explore the shifts needed for successful implementation. The efficiency of different AI and ML methods in various manufacturing practices will be evaluated [3, 7]. This research has the potential to greatly impact the future of knowledge and practice in manufacturing, leading to significant advancements in efficiency, quality, and innovation [2].

1.3. Research Objectives

Based on the summary of the entire essay, this section will outline the purpose of the study. The first objective is to understand what artificial intelligence and machine learning are and how they are used in manufacturing [4]. The second objective is to explore the different applications and uses of AI and ML in the industry and what benefits and challenges their adoption. The third objective is to provide real-world examples through a comparative analysis and analyze the differences between artificial intelligence and machine learning when applied to manufacturing [5, 8]. Finally, the last and main objective of the research is to identify which technology - either AI or ML - is best suited to different manufacturing processes to enhance efficiency and innovation. By achieving these objectives, this research can provide a guideline for manufacturing plant managers, innovation team managers, and technology strategy developers to understand better the current progress in artificial intelligence and machine learning in the industry and to make informed decisions on introducing these technologies into their manufacturing processes and systems (As shown in Figure 3).

Figure 3

Differences between AI and ML

The literature review establishes the importance and principles of the research, appraises previous studies, and is crucial for research projects and students' first hurdle. Write an extensive literature review using credible authors' secondary data. It tests critical thinking and communication skills [4]. It should have a clear purpose and critical analysis. Search for required information with a clear idea. Organize the literature thematically. Each part should have an introduction and critical comment. State the source and assessment. Conclude by summarizing the main theme.

2.1. Overview of AI and ML in Manufacturing

AI and ML are essential in the manufacturing industry, enabling tasks like understanding language, recognizing patterns, and learning from experience. Machine learning focuses on developing algorithms to improve computer performance for specific tasks. Major car manufacturers utilize AI and ML for driver-assistance systems. Mars, a global food company, invested in AI-powered procurement practices to modernize its supply chain and improve supplier collaboration. These examples demonstrate the practical applications and value of AI and ML in manufacturing.

2.2. Applications of AI and ML in Manufacturing

With tight control simulations, AI links sub-processes and imitates an actual production facility. This enables predicting outputs in different scenarios and enhancing performance. These computer programs, known as expert systems, solve complex problems using heuristic rules [3, 5].

AI enables sales and production planning, generating stock levels and reordering quantities based on coverage, savings, and service level. It estimates daily usage and production lead time, performing calculations efficiently for real-time execution and up-to-date guidance. AI's superiority in time operations allows for more simulation runs to make decisions on service and costs.

OEE is a benchmarking tool for measuring manufacturing effectiveness. AI is used to analyze OEE data and uncover patterns. This falls under predictive analytics in manufacturing (As shown in Figure 4).

Figure 4

Machines connected with sensors

ML is crucial for predictive maintenance in manufacturing. It uses data analysis tools to detect anomalies and prevent failures. It is the most effective technique compared to preventative and reactive methods.

AI and ML are used in manufacturing for design, distribution, and self-managed production. AI aids in generative design, generating shapes and structures efficiently. Engineers can explore optimized designs quickly, producing parametric designs [3].

2.3. Benefits and Challenges of and ML AI Adoption

This is a subchapter of the literature review where the benefits of AI and ML technology in the manufacturing industry are extensively discussed. However, I am bringing this section as a separate subchapter for the readers to understand the positive aspects of AI and ML adoption within manufacturing and, simultaneously, to comprehend the challenges the industries face during the AI and ML technology adoption period [7] (As shown in Figure 5).

Figure 5

Challenges of implementing AI in manufacturing

First, the focus will be on the benefits of using these technologies, and then the challenges will be discussed. The information in this section is crucial as it helps to direct the reader's mind so that when they read the comparative analysis, they will have in mind the positive aspects and the challenged state that technology is not perfect [6].

The chosen research design for this study is a case study comparing two organizations implementing artificial intelligence and machine learning in manufacturing processes. The data collection method includes document analysis of business and strategic documents, project charters, meeting minutes, and project management documents. A thorough analysis of the project's different aspects helps identify potential problems and solutions. AI and machine learning development and maintenance are conducted in a controlled environment, focusing on change control, risk management, and solution development processes.

3.1. Research Design

The methodology section explains the research design, data collection, and analysis methods used in the study. Descriptive research describes the characteristics of a population or phenomenon without manipulating variables. It helps researchers understand the current situation, what is essential to the population studied, and where more information is needed. The article discusses selecting a case study from the literature for data collection and analysis based on qualitative and quantitative methods. The aim is to help researchers choose a research type for their explorative studies. Critical theorists focus on creating a framework for less advantaged individuals, emphasizing equality and emancipation. Structured interviews, a quantitative method, are used throughout the article. A questionnaire is developed, and standard questions are used in a standard order to reduce method bias.

3.2. Data Collection

Machine learning and AI in manufacturing rely on data quality. The study requires primary data on shop-floor practices and the extent of AI and ML implementation. This includes machine-to-labor ratios, existing automation, industries adopting AI, and the impact on production chains. Data will be collected from a representative sample of Malaysia's production chains. Real-time timestamp data can be collected from a system recording machine progress. Secondary data will be collected through web surveys and literature reviews to understand AI in Malaysian manufacturing. This secondary data will help set the stage for primary data collection and understand the challenges. Qualitative and quantitative data collection methods will be used to maximize information and opinions. The research concludes that Malaysia will surpass other ASEAN countries in AI and ML implementation [4, 5] (As shown in Figure 6 and Figure 7).

Figure 6

AI Data collection methods

Figure 7

Mapping for Data Collection

The government should provide incentives to encourage the Industrial Revolution. Future research can be done through continuous investment in AI and ML technology in Malaysia.

3.3. Data Analysis

The OLS and Hausman tests were performed to test for endogeneity. The Hausman test rejected the null hypothesis of no endogeneity, indicating that our choice of IVs is relevant. 2SLS was used in this study. GrossmanModel1 regression model was formed using all variables, with AI as an exogenous variable. GrossmanModel2 was formed using AI as the dependent variable and different types of investments as independent variables. LM tests detected spatial correlation in the residuals. Only the results of 2SLS and Bayesian analysis will be reported [7]. Bayesian analysis showed a significant and positive coefficient for AI in GrossmanModel2, suggesting a positive impact of AI on health. Wilcoxon rank sum test revealed a significant difference in AI investment between states with low and high average life expectancy, confirming RQ4 (As shown in Figure 8).

Figure 8

Data Analysis for AI & ML Companies

Two conclusions are drawn from the analysis. Firstly, AI and ML effectively optimize machinery systems in manufacturing processes, as evidenced by the case study. This supports the selection of either method. Secondly, AI can significantly increase productivity and reduce work-in-process inventory. It enables real-time adjustment of production scheduling based on customized orders [5].

In contrast, ML-based CNC machines only address bottlenecks and lack flexibility in production scheduling. ML can predict and correct production irregularities, but process optimization and control progress is slow. Combining ML and other methodologies is necessary to achieve global optimization, methodology switching, and quality control updates for intelligent and autonomous production.

The comparison and evaluation between AI-utilizing robots and ML-utilizing CNC machines for end-product creation is based on product quality consistency and faulty parts. Table 1 indicates that CNC machines have a 50% defective rate, while AI machines have just 4.2%. AI-based robots produce nearly four times the daily amount of parts compared to CNC machines. However, more finished products are left to be delivered with CNC machines [7].

The analysis compares two case studies. Machine breakdowns decreased significantly in both cases. The breakdown percentage for the AI case decreased from 0.8% to 0.55% over the past four years. The breakdown trend in the ML case remained constant with a negligible slope. Hypotheses were tested using statistical methods. The results are summarized below.

4.1. Case Study 1: AI Implementation in Manufacturing

With KPIs set up to compare AI and human-run operations, the repair and maintenance system for machinery and equipment was then overhauled. By leveraging the historical data collected for all machinery and equipment spending, the AI developed predictive algorithms and was tasked to identify any potential issues before they even surfaced. So, how does the AI do this? Instead of grouping the machinery and equipment to be maintained periodically over the year like a human would do, the AI analyzes the machinery's performance over time from the data obtained from the sensors. These sensors could detect minute spikes or drops in energy consumption or output, vibrations, and abnormal sound emissions, which might not be picked up by routine scheduled maintenance. Hence, with predictive maintenance, plant and operation managers can better optimize the workforce, reduce maintenance duration, and reduce operation downtime, which translates to direct savings compared to the previous human-originated strategies [8] (As shown in Figure 9).

Figure 9

Comparative Analysis

4.2. Case Study 2: ML Implementation in Manufacturing

The "results" must be clearly stated, and a table can show a performance comparison. If the ML used in the project is shared, the repository link and conclusion should be included in this section. This evidence helps gain credit for extending programming knowledge by providing a practical solution and promoting an advanced programming tool.

ML algorithms are widely used, and sharing experiences in the programming society is vital. Including accurate codes in the appendices enhances the inspiration and credibility of the next section on "code and knowledge sharing."

The second heading is "Implementation and Testing". It explains how the proposed solution is introduced and the major tasks involved, such as converting existing data and creating learning materials. It also mentions how the testing is carried out (As shown in Figure 10).

Figure 10

Implementation in manufacturing

The third heading may be named "The Solution." In this section, a step-by-step procedure of the proposed ML solution can be outlined. This part requires a thorough analysis of each step, and a flow diagram showing the process should be included to make it more practical [9].

A brief introduction to the case study and background information are given. The first heading is "The Challenge," where the current setup or situation in the manufacturing company is described, possibly with a picture or diagram.

4.3. Comparison of AI and ML Approaches in Manufacturing

AI and ML are potent tools for manufacturers to optimize production and reduce costs. ML learns from outcomes, while AI focuses on algorithms. A challenge is the need for a better understanding of the digital journey. Data quality is crucial. AI can be a "black box," and ML still presents challenges. Technologies should align with business strategy. Adoption requires openness and support [7].

Clear project aims and evangelism are essential for success (As shown in Figure 11).

Figure 11

AI and ML Approaches in Manufacturing

When using AI or ML in a manufacturing environment, using knowledge graphs to organize the information could lead to a better understanding and supervision of the entire production line. This is because the links between the different stages of the production process are made explicit by the knowledge graph, which facilitates the discovery of bottlenecks or the identification of opportunities for optimization and innovation. Also, the transparency and interpretability of the results obtained from AI and ML models are always important, especially in a safety-critical domain such as manufacturing. Therefore, the research direction is to develop suitable and convincing explanations for the decisions and recommendations made by AI and ML models. This includes interpreting the models and their performance, exploring the accuracy, fairness, and any biases in the model's outcomes, and testing if a model can retain its robustness when data or model structures are subject to adversarial manipulations [10].

Moreover, since the amount of data that can be generated or used throughout the entire production process is increasing, the scalability and performance of AI and ML models become a concern. Therefore, another future work could be to adopt and tailor distributed systems and scalable algorithms for extensive data analysis, build intelligent systems capable of dealing with data diversity and volume in manufacturing, and offer real-time analysis and decision support. Also, given that implementing AI and ML in the manufacturing industry often involves a significant financial investment, a return of investment analysis, both in the short and long term, could provide valuable support to manufacturers and decision-makers [6].

This is also a research direction stated in future work because it is essential to validate the value of AI and ML before their large-scale adoption. The developed models may include cost estimations, a comparison analysis to traditional approaches, and case studies in different industrial sectors, which would be a fascinating research topic for technology and business and management studies [5].

This paper aims to compare AI and ML implementation in manufacturing. We discuss their deployment and significant changes. Data was collected from two companies and analyzed to determine future needs. Similar statements from Company "A" and "B" managers reveal common patterns. Conclusion: technological and strategic changes will impact every company. AI increases strategic goals and involves management in significant changes. The research addresses advancements in the field and suggests further investigation into robotics, 3D printing, and advanced analytics. Future researchers can establish causal relations between critical factors and work.