The Role of Artificial Intelligence in Sensory Evaluation of Food
Sensory evaluation serves as a cornerstone in food science, significantly influencing product development, quality assurance, and consumer preference research. It encompasses the measurement of essential sensory attributes—including taste, aroma, texture, and appearance—to align food products with market expectations (Bareen et al., 2025). Traditionally, sensory evaluation relied heavily on trained human panels and analytical instrumentation. While these expert panels bring valuable perceptual insights, they can be subjective, time-consuming, and necessitate extensive training. On the other hand, instrumental methods like gas chromatography and rheology offer objective measurements, though they might not fully capture the intricate, multisensory experiences of eating (Martinez-Velasco et al., 2024).
Advancements in Sensory Evaluation: Enter Artificial Intelligence
With the rise of artificial intelligence (AI), the scope of sensory evaluation is undergoing transformative changes. Tools such as machine learning (ML), computer vision (CV), natural language processing (NLP), and intelligent sensors are facilitating the high-throughput analysis of vast, diverse datasets that include chemical signals, image features, and consumer feedback (Chi et al., 2024). These advanced technologies are increasingly incorporated into sensory evaluation frameworks, yielding scalable and reproducible solutions that can potentially supplement or even replace traditional methods (Srivastava et al., 2025).
As illustrated in recent research trends, there has been a steady increase in publications focusing on “sensory evaluation using ML techniques” from 2015 to 2025. Deep learning, a specialized subset of ML, employs artificial neural networks (ANNs) to learn hierarchical features from large datasets, making it particularly suitable for high-dimensional inputs like images and spectroscopic data (Janiesch et al., 2021). Unlike traditional ML approaches that rely on manually selected features, deep learning excels at identifying complex patterns in sensory data such as aroma classification and flavor image recognition (Chang et al., 2021; Qiu et al., 2024).
The Shift Toward Data-Driven Approaches
The convergence of AI and sensory science signals a broader shift to data-driven methodologies. Keyword co-occurrence network analyses reveal that terms like "flavor prediction," "consumer preference," "e-nose," and "deep learning" regularly appear together, indicating the maturation of research around predictions and simulations relevant to the sensory experience. Reviews by Cosme et al. (2025) and Zatsu et al. (2024) delve into how digital sensory systems and virtual reality applications are revolutionizing the sensory evaluation landscape, highlighting the relevance of real-time consumer feedback modeling.
Notably, Cui et al. (2025) provide an in-depth examination of AI models utilized in food flavor development, showcasing how computational chemistry and taste databases can enhance sensory analysis. While existing literature often focuses on narrow applications—whether flavor modeling or digital sensors—this discussion provides a structured overview of interconnected AI applications in sensory evaluation, organized into key areas: ML for predictions, CV for texture and appearance, NLP for consumer insights, and intelligent sensors.
Transition from Traditional to AI-Based Methods
As AI technologies continue to evolve, the distinctions between traditional and AI-based sensory evaluation methods become clearer. Conventional approaches often rely on human judgment and fixed protocols, whereas AI systems incorporate sensor data, adaptive modeling, and multimodal inputs for more rapid and consistent assessments. This paradigm shift is enhancing the methods by which sensory data is gathered, analyzed, and applied in both industry and research contexts.
AI is not just about making predictions; it’s also being leveraged to simulate sensory experiences. For example, AI models can simulate taste and flavor profiles based on data from mass spectrometry and electronic noses, enabling quicker prototyping and significantly reducing the need for repeated human sensory testing (Queiroz et al., 2024).
Personalization in Sensory Experiences
The advent of customized sensory systems is transforming how we think about taste and preferences. AI can generate personalized flavor recommendations, optimizing product formulations for specific consumer groups based on individual data such as health status and cultural preferences (Vallabhaneni et al., 2024). This aligns with the broader trend in the food industry toward digital health and personalization. Nevertheless, challenges remain, particularly regarding the quality and variety of training data, as well as ethical considerations like data privacy and algorithmic bias (Agu et al., 2024).
Current Technologies and Future Directions
The present landscape of AI in sensory evaluation provides a rich foundation for exploring new technologies and methodologies. This includes investigating the core AI technologies relevant to sensory evaluation, exploring methods for simulating and predicting sensory attributes, and addressing both ethical and technical challenges associated with these advanced systems.
Notably, future directions emphasize the need for inclusive and scalable AI systems in food sensory science. Unlocking the full potential of AI technologies while mitigating their ethical risks will be vital for the responsible advancement of sensory evaluation practices. The continuous evolution in this field promises a more nuanced understanding of how we perceive food and how those perceptions can be harnessed for better consumer experiences.