We spoke with Andrey Aksenov about the specifics of applying artificial intelligence in mechanical engineering

Artificial intelligence is becoming increasingly integrated into our daily lives, transforming familiar processes and opening up new possibilities. The development of artificial intelligence unlocks the potential of business, science, medicine, education, and many other fields. The mechanical engineering industry is no exception. Andrei Aksenov discusses the symbiosis of artificial intelligence and mechanical engineering.

How did you enter the mechanical engineering industry and become a top manager?

My professional journey began at a machine-building enterprise in 2010, while I was still a university student. My first IT tasks focused on automating technical documentation and internal company processes. Over the course of several years, I headed a bureau and later the automation department. I succeeded in creating and implementing business process optimization methods that were innovative at the time, which led to increased efficiency and reduced production costs.

Now, one of the key areas of my work is the development of data-driven services, including those based on data from vehicle telematics devices. My experience in this field includes not only building infrastructure for data collection and analysis but also actively participating in the creation of unique services for clients. My work has not only facilitated the adoption of new technologies but also created value for clients by supporting them in their efforts to optimize operations and make informed business decisions based on data analysis.

Do you use artificial intelligence in your work?

Yes, working in a cutting-edge field, I always use new and modern technologies. My work involves developing and implementing solutions based on machine learning. Additionally, I regularly use modern AI-based services, such as GPT, to improve and streamline work processes.

What pros and cons can you identify regarding the implementation of artificial intelligence in the mechanical engineering industry?

When it comes to mechanical engineering, the implementation of artificial intelligence offers a number of significant advantages, but it also comes with certain challenges and risks.

I’ll start with the pros: artificial intelligence boosts productivity and efficiency; many processes can be automated, which shortens production time and reduces the likelihood of errors. Artificial intelligence allows for the optimization of workflows—it speeds up analysis and testing, leading to rapid innovation and lower development costs. Using AI to analyze sensor data allows us to predict equipment maintenance needs before serious problems arise, which minimizes downtime and reduces operating costs. Finally, artificial intelligence contributes to improved product quality: it analyzes production processes in real time, identifies and corrects deviations, which helps improve quality.

There are, of course, downsides: first, implementing AI systems requires significant investment in equipment, software, and staff training. Second, integrating artificial intelligence into existing production systems can be challenging due to the need to align it with existing equipment or software.

 Another significant challenge is the potential shortage of specialists capable of developing and maintaining complex AI systems in the labor market. I must also mention ethical and social issues—automation can lead to job losses.

Thus, despite significant advantages, the implementation of artificial intelligence in mechanical engineering requires careful planning, significant investment, and consideration of potential risks and challenges.

What specific skills and knowledge will employees need to successfully work with artificial intelligence systems in the mechanical engineering industry?

There are no special requirements in mechanical engineering; rather, they are standard for AI projects: an understanding of the principles of AI and machine learning, including supervised, unsupervised, and reinforcement learning; knowledge of the programming languages most commonly used in AI development, such as Python, including machine learning libraries (e.g., TensorFlow, PyTorch, Scikit-learn); the ability to work with large volumes of data, including its collection, cleaning, visualization, and interpretation using data analysis tools.

Such a specialist will also need in-depth knowledge of linear algebra, probability, statistics, and numerical methods, which are necessary for understanding and developing machine learning algorithms; the ability to approach problems analytically, critically evaluate data, and find effective solutions.

An important skill will be the ability to work in a team, communicate effectively with colleagues from different fields of expertise and with varying levels of technical training, as well as a willingness to engage in continuous self-learning and adapt to rapidly changing technologies and methods in the field of AI.

Do you think the implementation of artificial intelligence in mechanical engineering will affect cybersecurity?

In my opinion, the implementation of artificial intelligence in mechanical engineering could lead to the following threats: first, industrial control systems that integrate AI to improve efficiency and automation could become targets for cyberattacks. Vulnerabilities in these systems could lead to unauthorized access, which could potentially cause production shutdowns.

Second, AI systems in mechanical engineering rely on the accuracy and integrity of input data to make decisions. Data attacks, such as data poisoning, can distort AI conclusions, leading to erroneous actions or analysis.

Third, attacks targeting self-learning AI can deceive these systems, causing them to misinterpret input data. This can lead to equipment malfunctions or disruptions in production processes.

Finally, many AI projects in mechanical engineering may rely on third-party providers of cloud services, software, and components. This creates additional risks related to the security of the supplied solutions and the possibility of external attacks through the supply chain.

To minimize these threats, a comprehensive approach to cybersecurity is required, including thorough testing and auditing of AI systems, staff training, the development and implementation of incident response strategies, as well as collaboration with trusted suppliers and cybersecurity experts.