Researchers at Silesian University in Opava have made a surprising discovery: generative AI can help companies significantly reduce their carbon footprint. But there is a catch, it works only when the entire product life cycle changes.

Manufacturing accounts for 38% of global carbon dioxide emissions. Yet even as companies invest billions in ‘green’ initiatives, the industry’s carbon footprint continues to grow. In Pakistan, for example, air pollution claims 2.4 million lives every year.

But what if we could tackle this problem using the very technology that is currently transforming every industry generative artificial intelligence?

That is precisely the question posed by an international research team led by Syed Muhammad Shariq of the School of Business Administration at Silesian University. Published in the prestigious journal Business Strategy and the Environment, their study analysed 155 manufacturing companies and produced findings that could change the way we think about industrial sustainability.

The Three Faces of Generative AI

The researchers divided the use of generative AI in companies into three categories, each of which works differently.

The first is process automation. Imagine a factory where an AI system monitors production lines in real time, detects anomalies and optimises energy consumption. According to the study, this application of AI has the strongest positive effect on companies’ business intelligence, with a coefficient of 0.760 - an exceptionally high value in social science research.

The second category is cognitive engagement: AI’s ability to communicate with stakeholders across the supply chain, from raw-material suppliers and employees to customers. Here, too, the researchers confirmed a positive effect, although it was weaker.

The third category produced a surprise. Cognitive insight the ability of AI to analyse data and generate strategic insights was negatively associated with business intelligence. Why?

The researchers offer an explanation: AI’s analytical capabilities alone are not enough. Unless they are integrated into real-world processes and combined with human involvement, they remain little more than a technological artefact.

The Missing Link: The End of the Product Life Cycle

The study’s most intriguing finding concerns what the researchers call ‘end-of-life product management’ - the management of products once they reach the end of their useful life.

Intuitively, we might expect green innovation to lead directly to lower emissions. The data, however, tell a different story. The direct relationship between green product innovation and carbon performance was statistically insignificant, with a coefficient of just 0.029.

Does that mean green innovation does not work? Not quite.

The key lies in an intermediary. When companies implemented systems for managing products at the end of their useful life - processes for recycling, remanufacturing and reusing materials - the relationship changed dramatically. Green innovation led to better end-of-life management (coefficient: 0.579), which in turn produced a measurable reduction in emissions (coefficient: 0.599).

In other words, manufacturing ‘green’ products is not enough. The entire system must be designed so that products can be returned efficiently to circulation at the end of their useful life.

The Theory Behind the Findings

The researchers placed their findings within a theoretical framework known as Organizational Information Processing Theory (OIPT). This theory holds that an organisation’s success depends on the alignment between its capacity to process information and the level of uncertainty it faces.

In the context of the climate crisis, that uncertainty is immense. Companies must respond to changing regulations, consumer pressure, volatile raw-material prices and technological disruption. Generative AI can serve as a tool that dramatically increases a company’s capacity to process complex information - from supply-chain emissions data to forecasts of future trends.

But, and this is crucial, capacity alone is not enough. It must be linked to specific organisational processes that put it to use. This is why automation and cognitive engagement work, while analytical insight on its own does not.

What This Means in Practice

The study offers several concrete recommendations for managers.

First, investment in AI should focus primarily on process automation and tools that facilitate communication with stakeholders. Analytical tools are valuable only when they are integrated into real decision-making processes.

Second, green innovations must be designed from the outset with the end of the product’s useful life in mind. This means modular design, standardised materials and clear instructions for recycling and remanufacturing.

Third, success requires the integration of the entire supply chain, not only the forward flow from manufacturer to customer, but also the reverse flow from customer back to manufacturer.

Limitations and Future Directions

The study has its limitations. Its data come from Pakistani manufacturing companies, which limits the extent to which the findings can be generalised to other contexts. Although a sample of 155 companies is sufficient for statistical analysis, further research using larger samples would strengthen the findings.

The Silesian University researchers also propose several directions for future research: How do information-processing capabilities develop into dynamic capabilities? What role does leadership play in implementing these systems? And how do these relationships differ between the service sector and manufacturing? Only further research can provide the answers.

The Bigger Picture

At a time when the world is grappling with a climate crisis while undergoing an unprecedented technological transformation, the Silesian University study offers an important insight: technology alone will not solve anything. Generative AI is a powerful tool, but its contribution depends on how it is integrated into organisational processes and connected to the entire product life cycle.

The path to carbon neutrality does not run through individual ‘green’ products. It requires systemic change and artificial intelligence can act as a catalyst for that change. But only if we understand that the end of a product’s life is, in fact, the beginning of something new.