This month saw the first-ever AI Safety Summit at Bletchley Park. With it came similar themes of curiosity, hope and anxiety that make this unknown quantity so fascinating.
Innovation is not just an ambition; it’s a necessity. With consumers demanding better quality, safety, and sustainability, increasingly AI is looking to provide some of these solutions.
While AI is typically seen as a way to optimize supply chains, minimize waste, and enhance product safety, an unexpected area that AI is venturing into is flavour formulation.
Flavour formulation is both an art and a science, combining precise ingredients to create a distinct taste. Flavourists leverage knowledge of food chemistry and sensory perception to create a formula that captures the desired taste and aroma.
Flavourists use natural or synthetic ingredients, (often hundreds in a specific composition), to deliver the right flavour profile, while also considering flavour stability, potential interactions in the final product, processing effects, and flavour longevity.
AI’s capacity to analyse data, recognize patterns, and experiment with novel combinations opens a whole new range of possibilities for flavour innovation. AI technologies like Philyra can process enormous amounts of data at an unbelievable speed, potentially speeding up development.
Several food manufacturers and flavour houses have experimented with AI already. McCormick & Company’s collaboration with tech giant IBM, pioneering the use of AI in their flavour and food product development to Firmenich, using AI to create the world’s first AI-generated flavour of a ‘lightly grilled beef taste’ for plant-based meat alternatives, and Givaudan’s Taste & Wellbeing Advanced Tools for Modelling (ATOM) technology.
However, AI, like all innovations, does come with its own set of challenges and drawbacks.
The primary concern for using AI in food technology is the issue of reliance on data [5]. AI technologies, including those for flavour formulations, are inherently data-driven. This means the insights and solutions provided by AI are only as good as the data it’s trained on. With incomplete, unrepresentative data sets, the AI system may generate inaccurate or unexpected results. Without human interaction, unbiased data sampling will be needed to obtain reliable outcomes.
Flavourists bring invaluable experience and intuition. AI systems, while able to process a vast amount of data quickly and efficiently, lack the human touch and conscience which AI cannot account for.
So, therefore, should we instead view AI as a tool meant to augment human capabilities, not replace them? AI can analyse data at scale and detect patterns beyond human reach, but it’s the flavourists who provide the creative impulse and interpret AI’s suggestions in inventive ways. Technologies like AI could be used symbiotically, as a supportive role to human creativity, not as a replacement.
Whether sceptical of AI or not, it’s an undeniable fact that it’s being used increasingly all aspects of work and life. The adoption of AI in flavour innovation is opening new avenues for creating exciting, flavours while potentially accelerating the pace of innovation and reformulation.
But it’s crucial to note that, while AI can help devise flavour formulations, the work involved in implementing these in a real-world setting—from sourcing ingredients to meeting various local and global standards for food products are practical and incredibly difficult elements to any product development that need people.
While AI has shown great promise in redefining flavour formulation processes, the industry needs to balance this with potential downsides and mitigate these risks responsibly. Figuring out how AI can be used to help, rather than replace.