Today, it is becoming increasingly difficult to avoid discussions about the application of artificial intelligence (AI) in various fields. Since its popularity surged in 2023, the use of AI has only continued to grow. For some, AI proves helpful in completing personal tasks, such as planning a trip, while others aim to leverage this technology to simplify work assignments, optimize processes, which is particularly important in the industrial sector.

However, like any new technology, AI also faces significant resistance worldwide, because of its function to mimick human intelligence and essentially do the work on behalf of humans, which raises concerns. But as Prof. Dr. Andrius Dzedzickis, an applied artificial intelligence expert at VILNIUS TECH, explains that there is no need for people to fear the computers and robots one day conquering the world or taking their jobs.

"AI is an excellent tool for performing complex, repetitive, or highly precise tasks, making it especially valuable in industry. Therefore, it shouldn’t be resisted—it should be harnessed," says the scientist.
 

Prof. Dr. Dzedzickis from the VILNIUS TECH Faculty of Mechanics explains that AI is now widely applied across various industries, from machinery and equipment manufacturing to agriculture. Its main purpose is to optimize processes, increase efficiency, and reduce costs.

"In industry, artificial intelligence analyzes equipment loads and condition parameters, predicts potential failures before they occur, detects defects, optimizes the efficiency of production lines (thus reducing waste), manages inventory distribution, plans transportation routes, and forecasts the demand for goods or components.

AI is generally not applied in areas where technological limitations, regulatory barriers, ethical concerns, or low demand exist. The examples of that would be handicrafts or small-scale production of unique goods, such as traditional textiles," notes Dr. Dzedzickis, a professor in the Department of Mechatronics, Robotics, and Digital Manufacturing.

Moreover, AI helps tackle challenges and simplifies tasks that would take humans significantly more time to complete. For instace, it might take a person several days to analyze, categorize, and identify patterns among thousands of data points, while AI can complete such tasks in mere minutes. Additionally, the technology is less likely to make errors and eliminates the need for people to spend long hours on tedious tasks.

When applied in robotics, AI can enhance the accuracy of robotic movements. Last year, VILNIUS TECH researcher Dr. Marius Šumanas, for his dissertation, developed a method combining machine learning algorithms. This method adjusts robot movements to make them as precise and reliable as possible—a particularly critical factor in the industrial and medical sectors, where micrometer-level accuracy is essential.
 

Although AI takes over some human tasks, humanity is not at risk of being left without work. The technology operates within strictly defined environmental conditions. While it can generate ideas and analyze data, Prof. Dr. Dzedzickis points out that AI cannot replace human intuition, creativity, or strategic thinking.

"AI, like any other automation technology, is reshaping the job market. However, fears of mass unemployment are unfounded. AI may replace certain unskilled jobs, but it will also create new positions and opportunities. AI is a tool that enables humans to expand their capabilities and will serve as our assistant, just like other technologies, without replacing us," the professor explains.

Additionally, the operation of AI systems is continuously overseen by specialists in the respective field. Depending on the type of system, these experts monitor data collection and its processing to prevent incomplete or inaccurate data, test algorithm training processes, and analyze incorrect decisions made by the system.
 

When discussing AI, questions of its reliability often arise. This is partly due to the so-called "black box" phenomenon, where certain AI decisions cannot be explained or predicted. Sometimes, an AI program might even "hallucinate," providing incorrect results, so repeatedly relying on them would lead to flawed outcomes.

According to Prof. Dr. Dzedzickis, such issues are more common in general-purpose AI applications, which use large models and work with vast data sets. However, these limitations and unpredictable results are well known. Industrial AI, on the other hand, is developed and used differently.

"When we talk about forecasting and process optimization, decisions or recommendations must be verified, validated, and only then implemented. If we are discussing technological equipment management, the issues of black boxes or hallucinations are less relevant because different, less global models are often used. Black box solutions are primarily employed to protect intellectual property. Moreover, more reliable algorithms, tested with high-quality, limited data, are utilized.

For industry-specific AI algorithms, where system functionality may be critical, first extensive testing under various conditions is conducted, and only after statistically verifying the level of reliability are these algorithms deployed. For instance, in product quality control, the acceptable error rate might be 10 or 100 misclassified items out of 10,000 or more," explains the professor.
 

This question is frequently debated by AI critics and is a popular theme in both pop culture and conspiracy theories. However, the professor assures that such fears are unfounded.

"This is just another myth, similar to others humanity has faced since the First Industrial Revolution whenever a more complex or advanced technology emerged. I believe such myths stem partly from anthropomorphism—when we assign human traits to complex and unfamiliar objects or technologies, projecting our perspectives and thought processes onto them without considering their actual principles of operation.

AI, however, is a mathematical algorithm designed to perform specific tasks. It is not autonomous, has no emotions or feelings, and lacks any mystical ambitions to take over the world. The greater danger lies in how we, as humans, choose to apply it. Like any other tool, AI can be used to create or to destroy," concludes Prof. Dr. Andrius Dzedzickis from VILNIUS TECH.

Article prepared by Milda Mockūnaitė-Vitkienė, Project Manager for Internal Communications at VILNIUS TECH

Photograph by Aleksas Jaunius