From steam engines to assembly lines with conveyor belts and factory robots, the manufacturing sector has consistently been at the forefront of technological advancements. Artificial intelligence is poised to represent the next significant breakthrough, perhaps the most substantial yet. But how will this impact employment in the coming decade?
Applications include managing plants, suggesting equipment repairs, designing products, and assembling components. Manufacturing is already extensively automated, utilizing sensors, software, and computing networks to oversee the performance, data, pressure, and temperature of industrial machines and processes. This level of connectivity is crucial at facilities that can extend over vast areas.
“In a refinery or petrochemical facility, there can be thousands — or even tens of thousands — of instruments, equipment, and valves needed to manage 250,000 to 500,000 barrels of oil daily and convert that into gasoline,” highlights Jason Urso, chief technology officer at Honeywell’s software division.
Within the next decade, over 80 percent of manufacturing plants are expected to incorporate AI to assist in operating these “control systems” and resolving related issues, he anticipates. For example, if a machine produces an unusual sound, a factory worker can request the AI software to analyze that sound, summarize the associated problems, and suggest potential solutions, according to Urso.
Some manufacturers are already investing in this kind of AI. For instance, United States Steel Corporation has announced its intention to use generative AI software from Google to assist its employees with truck maintenance and parts ordering.
AI is also increasingly influencing product development. AI-enhanced software can enable automotive engineers to create multiple 3D car designs in minutes instead of days, claims Stephen Hooper, vice-president of software development, design, and manufacturing at Autodesk.
“You can create 3D designs of new vehicle styles in a fraction of the current time,” he states. “You can manage aspects like wheelbase and vehicle type, and the AI will generate hundreds, if not thousands, of alternatives.”
Hyundai has utilized Autodesk software to aid in the design of components for a prototype vehicle that can transform its wheels into legs for walking and climbing, potentially serving as a rescue vehicle.
While robots have long been employed for assembly in factories, the next generation will feature AI-driven “humanoid” robots that will work in tandem with humans. These robots will possess enough dexterity and learning abilities to perform tasks such as picking and categorizing items, experts believe.
Early iterations could be operational within the next five years, forecasts Geordie Rose, co-founder and CEO of Canadian startup Sanctuary AI, which aims to develop the first robots with “humanlike intelligence.” Its latest model, Phoenix, stands 5ft 7in tall, weighs 70kg, and is capable of walking at speeds of up to 5km/h. Humans operate it now, but Rose predicts that it will eventually replicate human memory, vision, hearing, and touch.
The demand for humanoid manufacturing robots is expected to be “significant,” according to a recent Goldman Sachs report — especially in the electric vehicle manufacturing sector.
“The central concept here is to create a machine that comprehends and acts upon the world like a human,” explains Rose. However, creating a machine that can respond like a human “is obviously much more complex than developing one that can perform a few human tasks.”
Sanctuary’s robot can already sort mechanical components at human speed, but even Rose admits that further advancements are necessary. “The question is, how much time it will take for our robots to transition from the lab to the manufacturing floor,” he remarks. “That’s a very challenging question to resolve.”
Ultimately, robots equipped with artificial general intelligence (AGI) — the same level of cognitive capability as a human — will be able to design and produce items, predicts Rose. “You could ask a sufficiently advanced AGI robot to create and manufacture a new battery.”
Jobs that may be lost include those of production-line workers, quality-control inspectors, and machine operators. Integrating AI into manufacturing robots — which do not require salary increases or go on strike — could potentially render millions of conventional manufacturing positions obsolete.
Pascual Restrepo, an associate professor at Boston University and a scholar of industrial robots, notes that non-AI robots have already displaced between 6 million and 9 million manufacturing jobs worldwide since the 1980s, including around 500,000 in the US.
Now, most experts predict that AI will further contribute to job losses in manufacturing. In a survey conducted last year by recruitment firm Nash Squared, technology leaders from around the globe estimated that 14 percent of roles in manufacturing and automotive sectors would be lost due to “automation” technologies, including AI, over the next five years.
Production-line staff, quality-control inspectors, and machinery operators appear to be the most vulnerable to being replaced by AI. Gabriele Eder, who oversees manufacturing, industrial and automotive sectors at Google Cloud in Germany, notes that in these roles, AI-driven machines and equipment can “frequently operate with superior precision and consistency than human workers,” requiring less human input during manufacturing operations.
“Our members are deeply concerned [about AI taking their jobs],” states Kan Matsuzaki, the assistant general secretary at IndustriALL, an international union representing over 50 million workers in the mining, energy, and manufacturing sectors. He also mentions that his members recognize the potential advantages of AI, such as enhancing safety in manufacturing.
Equipping manufacturing workers to work alongside AI could assist them in adapting and reducing job losses, but options may be limited. “When someone reaches around 55 years old . . . can they be retrained to become [an] AI machine . . . specialist, for instance?” Matsuzaki questions. “[It] is very challenging to accomplish.”
New job opportunities: machine monitors, robot programmers, digital champions, forensic AI scientists. However, some specialists anticipate that AI will generate more new positions in manufacturing than it removes. They argue that manufacturing firms prefer to hire rather than let go of employees—yet they face a global shortage of skilled workers in manufacturing.
Emerging AI-related roles in manufacturing will include overseeing AI machines, tracking their performance, programming robots, and collaborating in “cross-disciplinary teams” with expertise in both data science and manufacturing, experts predict. Simultaneously, traditional roles will evolve and become more technology-centric instead of being superseded by AI, according to Marie El Hoyek, a specialist in AI and industrial sectors at consulting firm McKinsey.
“Some manufacturing positions will need to change,” she remarks. “I envision that in the future, you would require digital champions who are core manufacturing personnel but can effectively communicate their needs in digital terms to the digital team, stating ‘this is what I need you to address.’”
AI will boost the demand for “forensic AI scientists,” typically with tech backgrounds, who evaluate AI system performance, says Cedrik Neike, the CEO of digital industries at the German tech firm Siemens. “[We] require experts who [can identify] failure points to fine-tune them,” he adds.
How extensively these AI technologies are implemented remains subject to discussion. “The crucial question is, who will profit from this AI?” Matsuzaki asks. “When you implement AI and automation robots in manufacturing environments . . . you could reduce your workforce, leading to increased productivity and profits . . . but there’s no benefit for the workers.”
Artificial intelligence can serve as a potent tool for training in manufacturing, as it enables virtual simulations, tailored programs, and performance evaluation with feedback. By considering the most probable scenarios workers might encounter, AI can integrate various factors to create realistic scenarios ranging from simple to highly complex, whether concerning plant conditions, machine upkeep, standard operations, or material considerations.
These AI resources can even utilize real-time performance metrics or equipment data to enable workers to practice tasks or skills, ranging from frequently used abilities to advanced problem-solving and teamwork required for tackling the most demanding situations.
Detroit-based startup DeepHow identifies a chance to leverage AI to expedite skills training for shopfloor and other highly technical trades workers. The company’s platform captures expertise and practical skills, leveraging AI, natural-language processing, computer vision, and knowledge mapping to transform this information into instructional training videos.
DeepHow’s AI Stephanie platform assesses a video of a skilled worker executing a complex task, recognizes the involved steps, and subsequently produces a detailed training video.
Sam Zheng, co-founder and CEO of DeepHow, points out that generating video training content has historically been expensive and time-intensive.
“However, implementing AI to produce video training material drastically enhances your video creation capabilities, simplifying the process of content development and enabling the production of new training videos—without the necessity of hiring costly film crews or staffing up with video content experts,” he states.
With a single click, AI incorporates advanced features such as transcribing and translating video material, allowing specialized skills knowledge to be documented and disseminated to all in a multilingual environment or across various countries.
“An additional advantage is that there’s no need for a professional videographer to divide content into sections, incorporate headings or notes, or include subtitles; let the AI handle everything for you,” he mentions.
Zheng emphasizes that current learners are not turning to PDFs and manuals; they prefer YouTube and video resources to observe someone perform a task and replicate that individual’s methods and techniques.
“In industrial environments, businesses that utilize AI-driven tools to create training videos can customize the experience to fit their employees’ unique learning requirements,” he notes.
For instance, if specific keywords or methods resonate with an audience, AI-driven tools can assist trainers in leveraging that. Another factor to consider is accessibility: AI makes training available for workers regardless of their primary language and ensures video training is accessible for employees who are hard of hearing or deaf — meeting workplace policies and legal requirements.
“The capacity to tailor training for each worker’s learning or performance is among the most compelling applications of AI in manufacturing,” explains Claudia Saran, KPMG’s national leader in industrial manufacturing.
She points out that AI can provide real-time insights into performance and develop training or coaching that focuses on those developmental areas while offering the worker essential feedback along the way.
“For example, personalized training can differ by subject and by the level of detail covered,” Saran adds.
She states that the ability to tailor training for each worker’s development or performance is one of the more appealing AI applications in the manufacturing sector.
“AI enhances other training and development methods and does not replace traditional training provided by colleagues, supervisors, and plant managers,” Saran remarks. “It can be a valuable addition to the workforce training toolkit, but it also necessitates careful oversight and significant input to be effective.”
Zheng mentions that one of the most challenging—but potentially most rewarding—benefits of using AI-powered training tools is the capacity to transfer “know-how.”
“Experienced senior workers develop and master specialized techniques that enhance speed, safety, and efficiency in their jobs,” he states. “This personal knowledge can be documented and shared with other workers, boosting an organization’s overall competitiveness.”
Mixed Feelings from Employees regarding AI in the Workplace
The fast-increasing popularity of ChatGPT and other generative AI applications has the chance to instigate a workplace transformation, yet its adoption also raises concerns among employees.
These findings came from a ResumeGenius survey of 1,000 employees, revealing that 69% of workers worry about job loss due to the rise of AI, and nearly three-quarters (74%) anticipate that AI technology will render human workers unnecessary.
The research indicated that IT, manufacturing, and healthcare are the sectors perceived as most vulnerable to being supplanted by AI technology.
In spite of these worries, 75% of survey participants expressed a positive sentiment towards using AI at work, while 21% felt neutral and merely 4% had a negative view.
Agata Szczepanek, a job search expert at Resume Genius, remarks that the increasing popularity of AI correlates with rising apprehensions regarding its implications, which is natural.
“Sometimes it goes too far—many individuals believe that AI will eliminate human employees, and that’s a significant misconception,” she states. “This scenario will never come to pass.”
She clarifies that while automation is unavoidable and AI continues to reshape the workplace, it’s humans who design, implement, and oversee machines.
“Numerous jobs require attributes that cannot be instructed or programmed,” she observes. “These include a profound comprehension of human emotions, intricate decision-making, empathy, and more.”
Although AI technology is likely to bring about changes in the labor market, Szczepanek asserts there’s no need to fear that human employees will one day become unnecessary.
Eilon Reshef, co-founder and chief product officer of Gong, concurs that there will always be a requirement for a human aspect concerning generative AI tools.
“Rather than replacing jobs, we prefer to consider generative AI as a means to enhance the tasks performed by humans,” he explains. “As generative AI tools evolve, we will likely see implementations that reduce some administrative work, analyze customer interactions and data, and deliver strategic recommendations based on a thorough understanding of customer nuances and attitudes.”
Reshef suggests that to remain competitive as generative AI enters various sectors, individuals should concentrate on the strategic skill sets that they have already been applying within their roles.
“Generative AI will persist in automating tasks and freeing up time for workers in diverse industries,” he notes. “It will become increasingly vital to excel in areas where generative AI has yet to develop, such as understanding nuance and strategy.”
He acknowledges that many employees are uncertain about how generative AI will influence their roles.
Organizations looking to adopt AI should inform employees about best practices for utilizing the technology and provide a clear explanation of how leaders intend to implement these tools to enhance existing tasks, according to Reshef.
Before implementing any kind of generative AI, leaders need to explore how it can be applied within their organization.
This requires evaluating which business areas can benefit from generative AI’s ability to automate tasks, ultimately saving time while maintaining quality and customer satisfaction.
According to Reshef, organizations should assess whether the use of generative AI can make business processes more efficient to improve performance during challenging economic times.
Cristina Fonseca, vice president of product at Zendesk, highlights that in customer experience (CX), AI is likely to automate most repetitive customer interactions, such as handling returns.
“However, this doesn’t mean that the roles of customer service agents will disappear,” she explains. “Instead, these roles will shift toward a more personalized approach, enabling agents to engage with customers more thoughtfully and emotionally.”
Fonseca believes that tools like ChatGPT will enhance workplace productivity, especially in the CX sector, where agents can offload repetitive and low-value tasks.
“Leaders should aim to use AI as a beneficial resource for employees, particularly as CX agent roles transition to focus more on supervisory duties,” she notes. “It’s essential that humans oversee AI to ensure its responsible and ethical use and minimize unique CX risks, ensuring a positive customer experience.”
Szczepanek emphasizes that the labor market is rapidly evolving, and staying flexible and adaptable is crucial.
“With the rise of AI-powered tools, managers need to communicate openly with their teams about their usage,” she advises. “Collectively, they can define best practices and maximize the benefits of AI technology in their environment.”
She believes that when implemented thoughtfully and ethically, AI can enhance productivity, create smoother workflows, and alleviate employee stress.
“In essence, it helps us to work more efficiently and quickly,” she continues. “However, there is a persistent risk that individuals might misuse AI to neglect their responsibilities. It’s also important to remember that we cannot fully rely on machines at all times.”
What Is AI in Manufacturing?
Numerous applications for AI exist in manufacturing, especially as industrial IoT and smart factories produce vast amounts of data every day. AI in manufacturing refers to employing machine learning (ML) and deep learning neural networks to refine manufacturing processes through superior data analysis and decision-making.
A frequently mentioned AI application in manufacturing is predictive maintenance. By leveraging AI on manufacturing data, organizations can better forecast and prevent equipment failures, thus minimizing costly downtime.
AI offers various other potential applications and advantages in manufacturing, including enhanced demand forecasting and reduced raw material waste. AI and manufacturing are naturally interconnected, given that industrial manufacturing environments necessitate collaboration between people and machines.
Why Does AI in Manufacturing Matter?
AI is integral to the notion of “Industry 4.0,” which emphasizes increased automation in manufacturing and the vast generation and sharing of data in these settings. AI and ML are crucial for organizations to harness the value embedded in the substantial data produced by manufacturing machinery. Utilizing AI for optimizing manufacturing processes can lead to cost reduction, improved safety, supply chain efficiencies, and a range of additional benefits.
Transformative Role of AI in Smart Manufacturing
Artificial Intelligence (AI) is transforming the manufacturing industry by boosting automation and operational effectiveness. The application of AI technologies in smart factories enables immediate data analysis, predictive maintenance, and enhanced decision-making processes. This section delves into the various roles of AI in manufacturing, highlighting its effects on automation and operational excellence.
Examples of Automation in Smart Factories
Predictive Maintenance: AI algorithms assess machine data to anticipate failures before they happen, thereby reducing downtime and maintenance expenses.
Quality Control: AI systems employ computer vision for real-time product inspection, ensuring high-quality standards are maintained autonomously.
Supply Chain Optimization: AI improves supply chain management by forecasting demand changes and optimizing inventory levels.
AI Training Courses for Smart Manufacturing
Workforce training is crucial for the effective adoption of AI technologies. There are various AI training programs that concentrate on:
Grasping the basics of AI and its applications within manufacturing.
Gaining practical experience with AI tools and platforms.
Cultivating skills in data analysis and machine learning tailored to manufacturing scenarios.
Challenges and Considerations
Despite the considerable advantages AI offers in manufacturing, several challenges need to be addressed:
Data Security: As manufacturing operations become increasingly interconnected, safeguarding sensitive data is vital. It is essential to implement strong cybersecurity protocols to defend against potential threats.
Technology Transfer: Closing the gap between academic research and practical use in manufacturing is essential. Collaboration between academic institutions and the industry can promote the successful application of AI technologies.
Conclusion
The incorporation of AI in manufacturing represents more than just a fleeting trend; it signifies a fundamental transformation towards more intelligent and efficient production processes. By harnessing AI technologies, manufacturers can enhance their flexibility, responsiveness, and competitiveness in the global marketplace. As the industry evolves, continuous research and development will be crucial in unlocking the complete potential of AI in smart manufacturing.
The intersection of artificial intelligence (AI) technologies and manufacturing is widely recognized. As one of the first sectors to embrace computer-based technology in the 1970s, manufacturing has emerged as a significant player in AI by the 21st century.
Manufacturers are undoubtedly investing heavily in AI. Estimates suggest that the global AI in manufacturing market valued at $3.2 billion in 2023 is expected to expand to $20.8 billion by 2028.
This growth is unsurprising, as manufacturers clearly acknowledge AI’s critical role in their transition to Industry 4.0, fostering highly efficient, interconnected, and intelligent manufacturing processes.
Although the applications of AI in manufacturing are boundless, here are some of the most intriguing use cases:
1. Safe, productive, and efficient operations
After decades of using robots, manufacturers are now beginning to implement ‘cobots’ on their production floors. Unlike traditional robots that require separate enclosures, cobots can work safely alongside human operators, assisting in part picking, machinery operation, performing various tasks, and even conducting quality inspections to enhance overall productivity and efficiency. Highly adaptable, cobots can carry out numerous functions, including gluing, welding, and greasing automotive components as well as picking and packaging finished goods. AI-powered machine vision is essential for making this feasible.
2. Intelligent, autonomous supply chains
Utilizing AI, machine learning (ML), and Big Data analytics, manufacturers can achieve fully automated continuous planning to maintain supply chain performance, even under volatile conditions with minimal human input. Industrial companies can also leverage AI agents to optimize the scheduling of complex manufacturing lines. These agents can evaluate various factors to determine the most efficient way to maximize output with minimal changeover costs to ensure timely product delivery.
3. Proactive, predictive maintenance
By employing AI to monitor and analyze data from equipment and shop floor operations, manufacturers can detect unusual patterns to forecast or even avert equipment failures. For instance, AI can analyze vibration, thermal imaging, and oil analysis data to evaluate machinery health. The insights derived from AI also allow manufacturers to effectively manage spare parts and consumables, providing accurate predictions of downtime that can influence production planning and related activities. The outcome is enhanced productivity, cost efficiencies, and improved equipment condition. Generative AI can contribute additional benefits by reviewing documents, such as maintenance logs and inspection reports, to provide actionable and precise information for troubleshooting and maintenance tasks.
4. Automate quality checks
AI significantly alters the landscape of testing and quality assurance. Image recognition technology can automatically identify equipment malfunctions and product flaws. For example, AI models trained on images of both acceptable and defective products can assess whether an item may need reworking or should be discarded or recycled. Moreover, AI’s analytical strengths can be applied to identify trends in production data, incident reports, and customer feedback to reveal areas needing improvement.
5. Design, develop, customize, and innovate products
Generative AI can revolutionize product development by analyzing market trends, pinpointing regulatory compliance changes, and summarizing product research and customer insights. Armed with this information, product designers can innovate and enhance items while ensuring compliance by comparing specifications against the necessary standards and regulations.
The algorithms can swiftly create innovative designs that surpass the abilities of conventional techniques. This enables manufacturers to enhance the product qualities that matter most to them — safety, performance, aesthetics, or even profitability. For instance, in 2019, General Motors applied generative design to create a lighter and stronger seat bracket for its electric vehicles. Additionally, by employing AI tools and simulation software, manufacturers can develop, test, and improve product designs without requiring physical prototypes; this reduces development time and costs while boosting product performance.
By automating mundane and time-consuming tasks, AI allows manufacturing employees to concentrate on more creative or complex roles. AI can also suggest next-best actions, helping workers to operate more efficiently and effectively. Unlike earlier robots, contemporary AI systems, integrated with sensors and wearable tech, can alert factory staff to any dangers present on the shop floor.
Overcoming the data hurdle for implementing AI in manufacturing
In spite of these opportunities and substantial investments, manufacturers struggle to fully harness AI’s benefits.
A survey of 3,000 organizations across various industries and regions revealed that only 10% reported obtaining significant financial benefits from AI. This aligns with findings from the Infosys Generative AI Radar – North America study, which noted that around 30% of large enterprises ($10 billion+) have established generative AI applications that deliver business value, whereas fewer than 10% of companies earning between $500 million and $10 billion have done so.
While manufacturers acknowledge the necessity of integrating AI into their business operations, they feel discouraged by the outcomes.
The World Economic Forum’s December 2022 white paper titled “Unlocking Value from Artificial Intelligence in Manufacturing” identifies six obstacles to AI implementation in the sector, including a disconnect between AI capabilities and operational requirements, a lack of explainable AI models, and the considerable customization needed across different manufacturing applications.
AI algorithms require training on vast datasets that are clean, precise, and unbiased to function effectively. Since this can be challenging for manufacturers, many businesses end up utilizing small, fragmented, inconsistent, or low-quality data, leading to less than optimal results. Even when substantial data is available, it might not be readily usable by AI models.
Therefore, before supplying training data to AI, manufacturers must ensure it is harmonized so that all individuals within the organization — across various functions, business units, and regions — can access the necessary data in a unified format. Additionally, the data should be organized so that AI-powered software can generate on-demand insights tailored for specific users, such as factory managers, quality inspectors, and senior management.
The positive aspect is that once manufacturers tackle the major challenges of AI deployment, they can revolutionize every element of their business, yielding numerous advantages.
The concept of a fully autonomous factory has long been a fascinating theme in speculative fiction. This factory would operate with minimal human presence, entirely managed by AI systems overseeing robotic assembly lines. However, this scenario is unlikely to represent how AI will actually be utilized in manufacturing in the foreseeable future.
A more realistic view of AI in manufacturing is one that involves a variety of applications for small, discrete systems managing particular manufacturing tasks. These systems will function largely on their own and react to external incidents with increasing intelligence and humanlike responses—ranging from a tool’s deterioration, an equipment failure, to a fire or natural disaster.
AI in manufacturing signifies machines’ ability to carry out tasks similar to humans—reacting to both internal and external events, and even foreseeing certain situations—autonomously. The machines have the capability to identify a worn tool or an unexpected occurrence, and they can adapt and circumvent the issue.
Historians trace human advancement from the Stone Age through the Bronze Age, Iron Age, and so forth, measuring progress based on our mastery over nature, materials, tools, and technologies. At present, humanity is in the Information Age, also referred to as the Silicon Age. In this technology-driven era, humans have augmented their capabilities through computers, gaining immense power over the natural world, enabling achievements that were unimaginable to previous generations.
As computer technology advances toward accomplishing tasks traditionally handled by humans, the development of AI has been a logical step forward. Individuals have different choices regarding the application of machine learning and AI. One strong aspect of AI is its ability to assist creative individuals in achieving more. It doesn’t outright replace people; rather, the best uses empower individuals to excel in their unique strengths—in manufacturing, this may involve producing a component or designing a product or part.
The focus is increasingly shifting to the cooperation between humans and robots. Contrary to the common belief that industrial robots are fully autonomous and “smart,” many of them still necessitate significant oversight. However, they are becoming more intelligent through AI advancements, enhancing the safety and efficiency of human-robot collaboration.
How has the role of AI in manufacturing changed over time?
Currently, the majority of AI utilized in the manufacturing sector is focused on measurement, nondestructive testing (NDT), and various other processes. AI is aiding in product design, although the actual fabrication stage is still at the initial phases of AI adoption. Many machine tools remain quite basic. While news about automated shop tooling circulates, a large number of factories worldwide still depend on outdated machinery that has only minimal digital or mechanical interfaces.
Modern fabrication systems are equipped with displays—human-computer interfaces and electronic sensors that monitor raw material supplies, system conditions, energy use, and many other factors. Operators can visualize their activities, either via a computer screen or directly on the machine. The path forward is becoming evident, as well as the possible ways AI can be integrated into manufacturing.
Short-term scenarios include real-time monitoring of the machining process and tracking status indicators like tool wear. These applications fall under the umbrella of “predictive maintenance.” This represents an obvious opportunity for AI: Algorithms analyze continuous data streams from sensors, revealing meaningful patterns and applying analytics to foresee potential issues, alerting maintenance teams to address them proactively. Internal sensors can detect ongoing actions, such as an acoustic sensor picking up sounds of belts or gears beginning to wear, or a sensor assessing tool wear. This information can be tied to an analytical model that predicts how much operational life remains for that tool.
On the shop floor, additive manufacturing is gaining prominence and has necessitated the incorporation of various new sensors to monitor conditions affecting materials and fabrication technologies that have only recently been widely adopted.
The current status of AI in manufacturing
AI facilitates significantly more accurate manufacturing process design, as well as diagnosing and resolving problems when defects arise during fabrication, through the use of a digital twin. A digital twin serves as an exact virtual representation of a physical part, machine tool, or the item being produced. It surpasses a conventional CAD model, serving as a precise digital likeness of the part and predicting its behavior in the case of a defect. (Defects are inherent to all parts, which leads to failure.) The use of AI is essential for implementing a digital twin in manufacturing process design and upkeep.
Many small and medium-sized enterprises (SMEs) are attempting to surpass their larger rivals by quickly embracing new machinery or technology. Providing these services sets them apart in the fabrication sector; however, some are adopting new tools and processes without the essential knowledge or experience. This lack of expertise could stem from either design or manufacturing; entering the realm of additive manufacturing can be particularly difficult due to this. In such cases, SMEs might have stronger motivations for integrating AI than larger corporations: employing smart systems that offer feedback and support for setup and operations could enable a small newcomer to secure a disruptive position in the market.
In essence, comprehensive engineering knowledge can be integrated into a manufacturing workflow. This means that tooling equipped with onboard AI can come with the expertise necessary for its installation, adoption, sensors, and analytics to identify operational and maintenance challenges. (These analytics often feature “unsupervised models,” which are designed to detect sensor feedback patterns not linked to known issues by identifying unusual or “incorrect” elements that require further examination.)
A concrete example of this idea is DRAMA (Digital Reconfigurable Additive Manufacturing facilities for Aerospace), a collaborative research initiative valued at £14.3 million ($19.4 million) that began in November 2017. Autodesk is part of a consortium collaborating with the Manufacturing Technology Centre (MTC) to develop a “digital learning factory.” The entire chain of the additive manufacturing process is being digitally replicated; the facility will be adaptable to meet various user demands and allow the testing of different hardware and software solutions. Developers are creating an additive manufacturing “knowledge base” to facilitate the adoption of technology and processes.
In the DRAMA project, Autodesk is pivotal in design, simulation, and optimization, fully considering the downstream manufacturing processes. Understanding how the manufacturing process affects each part is crucial information that can be automated and integrated into the design process through generative design, enabling the digital design to align more closely with the physical component.
This scenario presents a chance to effectively package a complete end-to-end workflow as a product for manufacturers. It could encompass everything from software and physical machinery in the factory to the digital twin of the machines, the ordering system that communicates data with the factory’s supply chain systems, and the analytics that oversee manufacturing methods and gather data as inputs progress through the system. Essentially, this results in the creation of “factory in a box” solutions.
Such a system would permit a manufacturer to analyze the part produced today, compare it with yesterday’s product, confirm that product quality assurance has been conducted, and evaluate the non-destructive testing (NDT) performed for each process on the production line. The feedback would provide the manufacturer with insights into the specific parameters used to produce those parts and highlight defect locations using sensor data.
The ideal vision of this process would entail loading materials on one end and receiving finished parts at the other. Human involvement would be necessary primarily for system maintenance, while much of the labor could eventually be handled by robots. However, currently, people are still responsible for designing, making decisions, overseeing manufacturing, and fulfilling various line functions. The system aids them in comprehending the true effects of their decisions.
The strength of AI largely stems from the capabilities of machine learning, neural networks, deep learning, and other self-organizing systems to learn from experience without requiring human input. These systems can swiftly identify significant patterns within large datasets that would be unmanageable for human analysts. Nonetheless, in today’s manufacturing landscape, human specialists predominantly guide AI application development, embedding their expertise from prior systems they’ve created. Human experts contribute their understanding of past events, including what has gone wrong and what has succeeded.
In time, autonomous AI will leverage this repository of expert knowledge, allowing a new employee in additive manufacturing to gain from operational insights as the AI evaluates onboard sensor data for preventive maintenance and process refinement. This represents an intermediate stage leading to innovations like self-correcting machines, where tools adapt to maintain performance as they wear out while suggesting the replacement of worn-out components.
AI applications extend beyond the fabrication process itself. From a factory-planning perspective, facility layout is influenced by numerous factors, including worker safety and process flow efficiency. It may necessitate the facility’s adaptability to accommodate a series of short-run initiatives or frequently shifting procedures.
Frequent alterations can result in unexpected space and material conflicts, which can subsequently lead to efficiency or safety concerns. However, such conflicts can be monitored and evaluated through the use of sensors, and AI can play a part in optimizing factory layouts.
Sensors gather data for immediate AI evaluation.
When integrating new technologies with significant uncertainty, such as additive manufacturing, a crucial measure is employing NDT after the component has been fabricated. Nondestructive testing can incur high costs, particularly when it involves capital equipment like CT scanners that assess the structural integrity of manufactured components. Machines equipped with sensors can connect to models developed from extensive datasets gathered from the manufacturing processes of specific parts.
Once sensor data is collected, it becomes feasible to create a machine-learning model that utilizes this data—for instance, to identify issues correlated with defects found in a CT scan. The sensor information can alert to potential defects without needing to CT-scan every part. Only those items flagged by the analytic model would undergo scanning instead of routinely checking all parts off the production line.
The operation can also track how personnel utilize the machinery. Manufacturing engineers often assume certain operational behaviors when designing equipment. Human observation may reveal additional steps being performed or certain steps being omitted. Sensors can accurately document this behavior for AI analysis.
AI is also capable of adjusting manufacturing methods and tools based on varying environmental conditions they might encounter. For instance, in additive-manufacturing technology, it has been discovered that some machines do not function as intended in particular regions. Humidity sensors in the factories have been utilized to monitor conditions, sometimes uncovering surprising findings. In one instance, humidity problems arose in a moisture-controlled environment due to someone leaving the door open to smoke outside.
To effectively leverage sensor data, it’s essential to create robust AI models. These models must be educated to comprehend what they observe in the data—identifying causes of problems, detecting these causes, and determining appropriate responses. Currently, machine-learning models can utilize sensor data to foresee issues and notify a human to troubleshoot. In the future, AI systems are expected to predict problems and respond to them in real time. Soon, AI models will be responsible for devising proactive strategies to prevent issues and enhance manufacturing processes.
Generative design
AI plays a significant role in generative design, a method in which a designer inputs a set of requirements for a project, and design software generates multiple variations. Recently, Autodesk has amassed substantial materials data for additive manufacturing and is employing that data to fuel a generative-design model. This prototype has a “grasp” of how material properties vary based on how the manufacturing process influences different features and geometries.
Generative design is a versatile optimization approach. Many conventional optimization methods tend to focus on broader strategies for part optimization. Generative-design algorithms, however, can be much more detailed, concentrating on specific features and applying knowledge of the mechanical attributes of those features derived from materials testing and partnerships with universities. While designs may be idealized, manufacturing occurs in the real world, where conditions may fluctuate. An effective generative-design algorithm incorporates this level of insight.
Generative design can produce an optimal design and specifications in software, subsequently distributing that design to multiple facilities equipped with compatible tooling. This allows smaller, geographically dispersed facilities to manufacture a wider array of parts. These facilities could be located close to where they are needed; a facility could produce aerospace components one day and then switch to another essential product the next day, reducing distribution and shipping expenses. This concept is increasingly significant in the automotive industry, for instance.
Flexible and reconfigurable processes and factory floors
AI can likewise be applied to enhance manufacturing processes and render them more adaptable and reconfigurable. The current demand can influence factory floor arrangements and generate processes for anticipated needs. Those models can then be utilized for comparative analysis. This evaluation will ascertain whether it is more advantageous to employ fewer large additive machines or a multitude of smaller ones, which may be less expensive and could be redirected to other projects if demand decreases. “What-if” analysis is a common use of AI.
Models will be employed to enhance both shop floor configuration and process sequencing. For instance, thermal treatment on an additive part can occur directly from the 3D printer. The material might arrive pre-tempered, or it may need to go through a retempering process, requiring an additional heat cycle. Engineers could simulate various scenarios to assess the necessary equipment for the facility; subcontracting parts of the process to a nearby company might be a more practical approach.
These AI tools could alter the business rationale for determining whether a factory should specialize in a single process or diversify its offerings. The latter option would increase the factory’s resilience. In the case of aerospace, an industry facing a decline, it might be possible for its manufacturing operations to pivot towards producing medical components as well.
Manufacturing and AI: Uses and advantages
Design, process enhancement, machine wear reduction, and energy consumption optimization are all fields where AI will make an impact in manufacturing. This transition is already in motion.
Machines are becoming smarter and more interconnected, both with each other and with the supply chain and broader business automation. The ideal scenario would involve materials being input and parts being output, with sensors tracking every stage in the chain. While people maintain process control, they might not need to work directly in the environment. This allows essential manufacturing resources and personnel to concentrate on innovation—developing new methods for designing and producing components—rather than engaging in repetitive tasks that can be automated.
As with any significant change, there has been some resistance to the adoption of AI. The knowledge and expertise needed for AI can be costly and hard to find; many manufacturers lack these capabilities internally. They view themselves as proficient in specialized areas, so to support the investment for innovation or process improvements, they require comprehensive evidence and may be reluctant to expand their operations.
This makes the concept of a “factory in a box” appealing to businesses. More companies, especially small and medium-sized enterprises (SMEs), can confidently implement a packaged end-to-end process where the software integrates smoothly with the tools, utilizing sensors and analytics for improvement. Incorporating digital twin capabilities, where engineers can simulate new manufacturing processes, also reduces the risk in decision-making.
Predictive maintenance is another crucial area for AI in manufacturing. This enables engineers to outfit factory machines with pretrained AI models that encompass the accumulated knowledge of that equipment. Based on machinery data, these models can identify new patterns of cause and effect discovered on-site to avert potential issues.
AI can also play a role in quality inspection, a process that generates extensive data, making it naturally suited for machine learning. Take additive manufacturing as an example: a single build can generate as much as a terabyte of data concerning how the machine produced the part, the conditions on-site, and any problems identified during the build. This data volume surpasses human capacity for analysis, but AI systems can manage it effectively. What is applicable for additive tools can similarly extend to subtractive manufacturing, casting, injection molding, and various other manufacturing techniques.
When complementary technologies such as virtual reality (VR) and augmented reality (AR) are integrated, AI solutions will shorten design time and streamline assembly-line operations. Workers on the line have already been equipped with VR/AR systems that allow them to visualize the assembly process, providing visual cues to enhance the speed and accuracy of their tasks. An operator might use AR glasses that display diagrams detailing how to assemble the components. The system can monitor the work and provide feedback to the worker: You’ve tightened this bolt sufficiently, you haven’t tightened it enough, or you’ve not pulled the trigger.
Larger corporations and SMEs have distinct priorities regarding AI adoption. SMEs typically produce numerous parts, while larger firms usually assemble many parts sourced from various suppliers. However, there are exceptions; for instance, automotive companies often perform spot-welding of the chassis while purchasing and assembling other components like bearings and plastic parts.
Concerning the parts themselves, a rising trend is the development of smart components: parts equipped with embedded sensors that monitor their own condition, stress, torque, and similar factors. This concept is particularly intriguing in auto manufacturing, as these elements are influenced more by how the vehicle is driven rather than the distance traveled; if consistently driven over rough terrain, more frequent maintenance will likely be necessary.
A smart component can alert a manufacturer when it has reached the end of its lifecycle or is due for an inspection. Instead of having to monitor these data points from the outside, the part itself will periodically communicate with AI systems to report its normal condition until something goes wrong, at which point the part will require attention. This method reduces the data traffic within the system, which can significantly hinder analytical processing capabilities at scale.
The most significant and immediate opportunity for AI to provide value lies in additive manufacturing. Additive processes are prime candidates because their products tend to be more costly and produced in smaller quantities. In the future, as humans develop and refine AI, it will probably become vital throughout the entire manufacturing value chain.
Data is shaping the future of manufacturing. The sector is undergoing rapid changes as significant trends and innovations transform how businesses operate in 2024 and beyond. Developments in robotics, artificial intelligence (AI), and the Internet of Things (IoT) are steering us toward more integrated, intelligent, and automated manufacturing solutions. This holds the promise of improved efficiency, lowered costs, and enhanced product quality.
According to Deloitte’s 2024 Manufacturing Industry Outlook, the remarkable growth in the manufacturing industry in 2023 can be attributed to three major legislative initiatives: the Infrastructure Investment and Jobs Act (IIJA), the Creating Helpful Incentives to Produce Semiconductors (CHIPS) and Science Act, and the Inflation Reduction Act (IRA).
Since these laws were passed, construction spending has experienced a significant rise, hitting $201 billion by mid-2023—a 70% increase from the prior year—thereby creating a higher demand for products. However, this growth comes with the combined challenges of geopolitical instability, skilled labor shortages, supply chain disruptions, and the necessity to meet net-zero emissions targets, requiring strategic adjustments.
Key Industry Trends
Tackling the skilled labor shortage is a top priority for us manufacturers. Adopting smart factory solutions could be a strong initial move to enhance productivity. Another essential area of focus is improving supply chain resilience through digitalization. The market has clearly indicated that excelling in customer service and aftermarket services is vital for staying competitive.
Kevin Stevick is the President and CEO of Steel Craft, a materials manufacturing company located in Hartford, WI.
Generative AI has considerable potential to transform several of these urgent challenges, particularly in product design, service quality, and supply chain management. Although still in its infancy, AI is expected to enable manufacturers to reduce costs and address labor issues.
1. Robotics and Automation
Collaborative robots (cobots) are gaining popularity, working alongside humans to boost productivity without displacing jobs. Designed for user-friendliness and safety in close human interaction, they fit well in tasks such as welding, assembly, and product inspection. A notable outcome is the reduction in lost time injury rates. They are also more affordable and versatile now, making it easier for SMEs to adopt previously unaffordable automation technologies.
2. AI
Importantly, AI is assisting in predicting maintenance requirements before equipment breakdowns occur. This can significantly reduce downtime and prolong the lifespan of machinery. AI-driven quality control, using advanced image recognition and machine learning techniques, makes it simpler for manufacturers to identify defects, minimize waste, and ensure superior product quality.
3. IoT Solutions
Central to the development of smart factories, interconnected devices are refining production processes through real-time data sharing. IoT is also enhancing supply chains by offering real-time tracking of products and enabling more efficient management by manufacturers. The advantages include lowered inventory costs and quicker adaptation to market changes.
Considerations for Testing the Waters
My organization, Steel Craft, is currently working to integrate more robotics and automation into our laser-cutting and brake press operations to boost our lights-out capability. I’ve realized that regardless of how beneficial technology might be, maintaining a stable workforce remains essential, tying back to an improved employee experience. This could involve revamping the benefits program or launching a bonus scheme.
Being proactive in implementing AI and robotics not only on the manufacturing floor but also in back-office processes can enhance your organization’s efficiency. As you train your staff to operate new automated equipment and support their transition from manual roles to more technology-driven positions, assuring employees about job security and benefits is critical.
By concentrating on data, manufacturing firms can position themselves in alignment with the latest industry standards, which is crucial to remain competitive and effective in today’s marketplace. We’ve noticed significant changes in our design and engineering processes since adopting computer-aided design and engineering software. Previously, we hadn’t fully harnessed the potential of data analytics. Incorporating these elements into our operations and shifting towards a data-driven approach has equipped us with the insights needed to inform decisions and refine our strategies.
I believe that merging traditional manufacturing with cutting-edge technology will allow the industry to maintain its growth momentum. It’s an exciting time for both the sector and its workforce. For successful AI integration, leaders need to engage directly with team members on the ground—the skilled workers on the shop floor and the specialists in the back office.
Recognizing repetitive and time-consuming tasks that can be automated is crucial for alleviating strain on employees, which in turn helps reduce feelings of burnout. As organizations continue to assign more mundane responsibilities to machinery and automation technologies, it becomes increasingly vital to invest in upskilling and cross-training initiatives. These programs not only equip employees with new skills but also open up a range of growth opportunities, enabling them to take on more complex and engaging roles.
Moreover, fostering motivation among team members is key to fully utilizing their expertise. When employees feel valued and empowered, they contribute more effectively, leading to enhanced collaboration between human workers and automated systems. This synergy not only improves overall operational efficiency but also elevates the quality of work produced. By focusing on both automation and employee development, companies can enhance productivity while ensuring that their workforce remains engaged and satisfied.
Leave a Reply