For decades, food quality assessments have been dominated by sensory panels and empirical indices. But as climate-driven shifts alter crop chemistry, industrial batches scale to thousands of tons, and consumers demand transparency around origin, authenticity, and sustainability, traditional quality tools are being stretched beyond their limits. 

A new generation of analytical approaches – from comprehensive two-dimensional gas chromatography to AI-driven “chemical imaging” – promise to make food quality predictive and actionable. But technology alone is not enough, argues Chiara Codero, Associate Professor of Analytical Chemistry at the University of Turin, Italy. Real progress depends on overcoming cultural resistance and rethinking how the food sector defines and values quality.

Here, Cordero explores how multidimensional separations and computer-vision strategies are reshaping the analytical foundations of food quality. She also considers the practical and structural challenges that stand in the way of broader adoption – and what must change for high-resolution chemical insight to translate into meaningful action across the food sector.

Could you take us through some of the key developments, tools, or turning points that have shaped food analysis in recent years? 

Food analysis has undergone a profound transformation over the past decade, driven by both instrumental and conceptual advances. Among the most significant is the full maturation of comprehensive two-dimensional gas chromatography (GC×GC). For years, the analytical community recognized the enormous potential of volatilomics, but the technology simply wasn’t yet ready to deliver the required level of stability, reproducibility, and resolution. The introduction of cryogen-free modulators and robust microfluidic flow modulators has changed this landscape entirely. These platforms have democratized access to high-performance multidimensional separations, allowing laboratories to achieve consistent, high-quality chromatographic fingerprints aligned with the true complexity of food matrices. 

Equally transformative has been the integration of AI and computer-vision approaches into chromatographic data interpretation. GC×GC inherently produces vast, richly structured datasets, and AI has provided the analytical bandwidth to interpret chromatograms as genuine chemical images. This shift – from peak lists to image-based fingerprints – enables deeper pattern recognition, enhanced classification accuracy, and a level of interpretability that aligns far more closely with human sensory perception. 

Beyond these technical breakthroughs, I believe the most important turning point is more conceptual: the recognition that food quality urgently needs innovation. Food safety experienced rapid evolution as a result of regulatory demand, but quality assessment still relies heavily on empirical, experience-based evaluations – such as sensory analysis, visual inspection and morphological analysis. Without new analytical strategies, we remain bound to subjective metrics that cannot capture emerging challenges such as climate-driven variability, new processing technologies, or evolving consumer expectations. 

In short, the last decade has provided us with the tools to act. Now, it is our opportunity – and responsibility – to use them to redefine how we measure, interpret, and ultimately value food quality. Better measurement is not just a scientific achievement; it is a societal imperative. 

Can you tell us about the focus of your current research or work in this area?

My group’s current research focuses on what we call “AI smelling” – a new generation of analytical workflows designed to capture, interpret, and translate the chemical information underlying sensory perception. The idea is simple, but powerful: integrate high-dimensional volatilomics and metabolomics with computer-vision tools to design an analytical platform which can interpret chemical signals with a level of nuance that approaches human sensory evaluation. 

Volatile molecules are often the earliest and most sensitive indicators of agricultural conditions, processing history, storage behavior, and authenticity. They reflect crop performance, environmental stress, microbiological activity, and the emergence of quality defects. By fusing volatile and non-volatile pathways, we gain a far more coherent and predictive picture of food behavior than any traditional empirical index can provide. 

A central part of our work focuses on hazelnuts – a complex, climate-sensitive, economically strategic crop. Spanning over a decade, we have developed GC×GC workflows, AI-based classification models, and multimodal data-fusion approaches to understand defect formation and climate-driven sensory changes, providing processors and growers with actionable markers long before defects become perceptible to human panels. This research has demonstrated that analytical tools can anticipate sensory deterioration, highlight early oxidation pathways and map positive aroma traits across large industrial batches.

Our broader mission is to help the industry shift from intuition-driven quality control to evidence-based decision-making. By identifying robust, chemistry-driven markers, we support more precise screening, reduce waste, and improve safety and consistency. Ultimately, our goal is to bridge analytical chemistry and sensory science, and to create workflows that do not replace sensory panels, but empower them – offering objective, high-resolution insight into complex food systems. 

Which analytical techniques or innovations are you most excited about right now? 

GC×GC remains one of my greatest scientific passions, especially as we learn to interpret its output through the lens of computer vision. When you treat a chromatogram not as a sequence of peaks, but as a complex landscape – with regions, textures, shapes, and intensities – you unlock relationships that aren’t visible using conventional peak-based analysis. Machine-learning algorithms trained on these image-like representations can identify subtle patterns, classify samples with great accuracy, and reveal chemical signatures that align well with quality attributes. 

Another area that’s gaining momentum is robotic sample preparation – essential for high-throughput, high-reproducibility workflows in volatilomics. Food matrices are notoriously complex and variable, and robotics help reduce the manual variability that often obscures minor – yet meaningful – chemical differences. This reproducibility is crucial when dealing with large-scale industrial challenges or longitudinal studies which track climate impacts. 

On the detection side, modern hyphenated MS instruments now operate at speeds and dimensionalities compatible with GC×GC, enabling more accurate, time-aligned, and confident identification. Add to this the power of data-fusion workflows, where volatilomics, metabolomics, sensory data, and biological endpoints are integrated into a cohesive interpretive framework, and we begin to see food systems as interconnected networks rather than independent signals. 

One recent example illustrates this evolution. We applied our AI-smelling workflow to a multi-ton hazelnut batch. Consisting of literal mountains of material, traditional sensory or chemical screening is unable to feasibly capture heterogeneity at this scale. And yet, in a single GC×GC–AI run, the system mapped positive odorants, detected faint rancidity markers, and highlighted early oxidative pathways. A quality manager standing beside me remarked, “It’s like having a sensory panel that doesn’t get tired.” 

That moment underscored how profoundly analytical decision-making can change when high-information tools meet robust automation and intelligent algorithms. 

What are the main technical or conceptual challenges still facing food analysis today?

Food analysis continues to face challenges that have persisted for decades. The intrinsic complexity of food matrices – chemically dense, dynamic, and highly variable – remains a central difficulty. Translating laboratory methods into industrial workflows is equally challenging: high-resolution techniques such as GC×GC are powerful but often require expertise, infrastructure, and training that some companies are hesitant to invest in. Standardization is another persistent gap; without harmonized protocols, cross-laboratory comparability and regulatory acceptance are slow to develop. 

Big-data interpretation has also become a critical bottleneck. Modern omics tools produce large volumes of information, and many laboratories lack the computational or statistical skills required to transform this data into meaningful decisions. While AI offers transformative potential, it also requires new competencies and cultural openness – attributes that are not uniformly distributed across the food sector. 

Yet in my view, the most significant barrier at the moment is cultural. Many laboratories – and even entire industries – are cautious when it comes to adopting unfamiliar technologies. If the advantages are not immediately tangible or quantifiable, enthusiasm quickly fades. There is an expectation that innovation should be effortless, that new tools should slot seamlessly into old workflows without conceptual investment. This mindset stands in stark contrast to the complexity of modern food systems and the urgency of emerging challenges like climate variability and sustainability.

I often ask: Do we really want to wait for regulations to force innovation, just as happened with food safety? Analytical chemistry has the capability to transform food quality evaluation, but only if we proactively embrace change. Overcoming cultural resistance is just as crucial as solving technical hurdles.

Progress requires not only better tools, but also a shared willingness to rethink how we measure, define, and value food quality. 

How are advances in data science, automation, or omics changing the way we study food quality and safety? 

Data science, automation, and omics technologies are converging to reshape how we understand and manage food quality. Instead of treating foods as static entities, these tools allow us to study them as dynamic systems whose chemistry responds to climate, processing, storage and biological interactions. Volatilomics and metabolomics capture early chemical signals reflecting crop stress, microbial shifts, or oxidative pathways long before conventional indices detect change. Automation ensures reproducibility and scalability, enabling routine, large-scale monitoring aligned with industrial demands.

One of the most important shifts is predictive capability. When climatic variations alter crop phenotypes, resulting changes in flavor, nutritional quality, or shelf-life may go unnoticed until they become problematic. With advanced analytical tools, we can detect these signals in advance. In our hazelnut research, volatilomics revealed climate-driven quality changes well before sensory panels or classical assays noticed them. This early warning enables growers and processors to adapt cultivation practices, adjust storage conditions, or industry redesign processing strategies. 

The societal implications are substantial. Improved analytical capacity supports sustainability by reducing waste through earlier detection of defects and better process control. It also enhances transparency and authenticity, helping to combat fraud and ensure claims about origin, quality, or purity are chemically substantiated. It supports nutritional quality by identifying how environmental or technological shifts influence bioactive compounds. And critically, it strengthens resilience: as climate change increases variability and unpredictability, our ability to monitor, interpret, and anticipate chemical responses becomes essential to maintain a stable food supply. 

In essence, omics, automation, and data science are not just analytical tools; they are strategic instruments for designing a more sustainable, transparent, and adaptive food system. 

Looking towards the next decade, where do you see the field heading? 

When I envision the next decade, I think less about specific instruments and more about cultural transformation. Technological breakthroughs will continue at a rapid pace – faster detectors, smarter algorithms, more integrated omics – but their impact will depend on whether we build a scientific ecosystem capable of using them effectively. 

I hope to see analytical chemistry recognized not as a service provider, but as a driver of strategic innovation. This requires a shift toward genuine transdisciplinarity, where chemists collaborate seamlessly with agronomists, food technologists, data scientists, engineers, sensory experts, and industry partners. Food systems are inherently interconnected, and solving emerging challenges – from climate-driven variability to sustainable processing – requires integrated expertise more than isolated interventions. 

Academia plays an important role here. As pressure grows to pursue fast, entrepreneurial outcomes, there is a risk that the slower but essential processes of mentorship, training, and deep conceptual thinking are diminished. I hope the next decade brings renewed commitment to nurturing curiosity, critical reasoning, and methodological rigor. These are the foundations of scientific longevity and innovation. 

If I could choose one breakthrough, it would be the widespread adoption of decision-making frameworks that integrate analytical data, sensory responses, biological insights, and sustainability metrics into a unified system. Such frameworks could guide agricultural choices, processing strategies, and policy decisions with unprecedented clarity. 

Ultimately, I envision a future where food analysis is valued not only for troubleshooting defects but for shaping healthier, more sustainable, and resilient food systems. In that future, curiosity, humility, and long-term vision are as important as any technological advancement. 

Newsletters

Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.

I have read and understand the Privacy Notice*

Subscribe