Would you trust an AI system to decide what you eat? A new app has been designed to analyze data and predict consumer preferences using machine learning.
Analytical Flavour Systems is a startup based in New York that has developed a tool that uses artificial intelligence to create personalized food based on your unique preferences.
The company, led by CEO Jason Cohen, is currently seeking $2 million in funding. Things are looking promising as it has attracted investment from industry giants such as TechStars, ZX Ventures, and AB InBev’s global innovation fund.
The company’s idea is based on the idea that the food industry supplies products aimed to please the masses. The result is an endless amount of food products that delight the few.
Analytical Flavour Systems aims to bridge this gap with their new tool called Gastrograph. The tool was developed on the basic premise that every person experiences flavor differently. That’s hardly groundbreaking. However, the tool uses machine learning to work out patterns in preference based on demographics and disarray in consensus.
Cohen told the Guardian that his AI tool will open the doors to manufacturing products that respond to customer’s individualized taste buds and preferences.
What’s Artificial Intelligence?
Artificial intelligence (AI) can be described as a robotic brain that is capable of gathering and analyzing vast amounts of data. This is efficiently carried out using machine learning or computational algorithms that can find patterns and make predictions. In other words, AI can improve itself by learning from the data it examines.
A robotic brain may sound futuristic, but you’ve probably already come across AI in your everyday life. It’s used to filter spam out of your inbox and to recommend what movies you’ll like on Netflix.
The Tech Behind the Taste Prediction App
Algorithms can’t taste things. How can a robot brain predict what a human will want to consume?
Cohen first began developing the product at Penn State’s College of Information Sciences and Technology in 2010. It started off as a basic AI platform that could tell you what tea you were drinking, its origins, and flavor notes. This platform eventually developed into Gastrograph AI.
Gastrograph AI is a smartphone app that features a 24 spoke wheel. Each section of the graph represents a category of sensory experience relating to food. For example “bitter” and “mouthfeel”. The user maps their perception of the flavor into this wheel by tracing the spokes corresponding to their qualities. The intensity of each is also recorded on a scale of one to five. Then, the taster gives the food an overall rating between one and seven.
The app combines this information with data it gathers about the user. This ranges from demographic information, past experience with the product, socioeconomic status, and smoking habits.
The app functions on the fact that our preferences for different tastes and textures are highly individual and shaped by biology, culture, and personal history. It also collects information about the environment the taster is in. The microphone, GPS, and light meter are all used during the process. Temperature, barometric pressure, and noise levels all have an impact on our experience of taste.
Gastrograph AI then uses this data to model and predict flavor preferences of certain groups of the public.
Why? To provide food and beverage companies with the insight they need to develop products optimized for different target audiences.
Sensory Science: the Measure of Pleasure
As it stands, a sensory panel decides what we eat. A trained group of specialists gets to decide everything from how crunchy your Doritos should be, to the creaminess of your cookie dough ice-cream.
Sensory panel tasters are trained to taste and smell food analytically, without any personal input. They use standardized vocabulary to describe aroma, texture, and taste.
Over the years, devices such as electronic noses and robotic tongues have been developed in an attempt to automate this important job. None so far have been successful, but Cohen believes his app could do a better job than the sensory panel.
One major problem with sensory science is that the majority of new products fail. This costs the US food industry $20 billion a year. Analytical Flavour Systems claim that their app can provide more reliable predictions when it comes to success rates.
Cohen claims that excluding personal preference from tasting panels is a huge mistake. Instead of narrowing down descriptions and experiences of food, the new technology can map popular opinion into the complex patterns that, until now, sensory science has largely ignored.
Possible Application of AI in the Food Industry
So far, Gastrograph has been used for quality control in the craft beer industry.
However, besides having an IPA personalized to your taste buds, AI has other possible applications in the food industry. There has been much buzz surrounding the use of AI to produce food to our metabolic needs and other forms of personalized nutrition.
With AI being used more widely in the food industry, we could be on the horizon of custom-tailored food. For example, Zipongo is a company that offers a digital nutrition platform that provides users with personalized meal recommendations based on biometrics and food preferences.
Habit is another personalized nutrition platform that uses machine learning with a similar vision. The founders see a reimagined food industry where DNA analysis, artificial intelligence, and food databases are combined to produce tailor-made nutritional plans.
The forthcoming possibilities of artificial intelligence merge into almost every industry. Gastrograph has the capacity to reveal patterns and preference on a large scale to provide a better experience with food and increase the success rates of new products. All AI needs in exchange is your data.