By Sai Sreenivas Kodur, Co-Founder & CTO
AI – A Brief History
To many computer scientists from the ’90s, AI was something which everyone tried but didn’t really work. But now, AI is rapidly changing the world, changing the way we interact, be it online or offline. It is not only inching towards the accuracy of humans (e.g. language translation, speech recognition, OCR) but also surpassing humans (e.g. object recognition, self-driving cars).
Until now, AI has largely been about pre-programming tools for specific tasks. In these kinds of systems, the intelligence of the system lies mostly in the smartness of the human who programmed all of the intelligence into the smart system. But, recently Google’s AlphaGO has gained much popularity in the field of general-purpose learning by combining techniques of deep learning and reinforcement learning, which is the closest attempt to replicate learning like humans by interacting with the surrounding environment. This opens up doors to replace a humongous number of tasks where the human level of expertise is required in executing daily tasks (e.g. car driving, security). Although, at the same time general purpose learning cannot be directly used in many tasks which involve creativity (e.g. movie making, literature), human experiences (e.g. travel, food), and others which are not solely based on intelligence or intuition. There is still a need for building custom smart solutions for a given problem.
Food & Culture – A Brief History
I’m sure all of us can recall instances when we attribute an old tradition or custom or story for a dish that is being served. It is indeed true, food has a long history and culture. The way food and recipes evolved is very comparable to that of natural language. Like language, food has a very intricate relation with culture and geography (more info on food and culture is available here). Going back in time before the spice trade (which led to first agrarian societies), food recipes were heavily dependent on the regional factors like:
– Ingredients available
– Cooking techniques known to people
This uniqueness of regional influence is the basis for the origin of cuisine which explains why every region has a different identity when it comes to food. Also, this is the reason why people from different regions tend to have different sensibilities towards food.
After trade started, food has become more global where ingredients from one part of the world are used in other places. This allowed new fusion cuisines that combine elements of different culinary traditions. This led to the creation of hybrid dishes, often the sign of a good modern chef.
Food Choices Are Personal
Food preferences of people are different, for example, some like it spicy, some like it sweet, some like that extra cheese in pasta, and many more. Everyone has a unique taste and experience the same food in different ways, for example, grilled fish is enjoyed by some and repelled by others (maybe because of smell and meat preferences). Human tastes have evolved over time, very similar to the evolution of life. Take a look at this article to understand how every animal has different sensibilities; it explains why hummingbirds are sugar junkies, while robins stick with worms.
Why people have different preferences for food?
– Similar to our varied preferences in non-food parts of life e.g. fashion, movies, music, etc., our food tastes are also different. Most of it is influenced by our culture and environment. People exposed to very similar conditions could also end up having different tastes, for example, siblings might have different tastes, even though they have grown up in the same conditions.
– Evolution of human tastes in food mainly happened because of our choices at every stage of life. Few tastes added up. Few did not.
– Globalization of food and hence eating habits, increase in choices and diversity of food was only possible because of globalization. For example, people in India who are exposed to majorly Indian cuisines during their childhood ended up liking or disliking new dishes which were exposed to them. Also in India, ingredients which were available to the people a while back were very limited compared to the current scenario. The purchasing power of individuals also plays a factor, some ingredients are expensive (example saffron, tuna fish, etc.), some are not (example chicken, common vegetables, etc.). This allows people with higher purchasing power to get exposed to a better gourmet experience. This is true in its own way, in each country as a whole, and also across countries.
– New dishes came in the way the science of cooking defined it. Few dishes came in experimentally. Modern chefs create new recipes in a well-defined fashion (example here) and hence new cuisines also evolved just like the way new genres evolved in music.
Applications of AI & Personalisation Technology in the Food Sector
a) Restaurants – A restaurant’s main agenda is to keep customers satisfied by providing them the highest quality of food and service and to operate profitably by minimizing costs. Technology played an important role in transforming this industry. Online presence through discovery, table booking, and food ordering opened up channels to attract new customers and keep old customers happy. While restaurant management and POS help improve the internal process for payment, accounts, CRM, inventory management. For any typical restaurant, there are two types of customers:
– Dine in: Service, convenience, ambiance, and food taste, food presentation are important factors.
– Food quality, packaging, delivery time are important factors.
Few additional problems that could be addressed by AI assisted technology:
– Long menus: Average of 120 dishes on a menu in India. Extremely hard to really go through every dish for a customer. Presenting a relevant menu gives the restaurant a chance to win a customer.
– Missing dish info: Lack of knowledge about the dish from the customer’s perspective.
– A restaurant has very little or zero knowledge about customer’s taste preferences, budget, dietary and health conditions. The restaurant’s kitchen could tap into data on customer’s food preferences and could serve better options.
– CRM built on context, food and dining preferences of a user. Targeting the right customers improves loyalty and deepens the relationship. Here are a few use cases:
i) Many existing and new restaurants make changes in their menus to make themselves relevant and fresh to customers, currently, there is no way to predictively measure the impact on their business.
ii) Reduce dependency on discounts in generating business.
iii) Negative feedback on the food is understood deeply.
b) Health & fitness – Food choices each day affect your health, how you feel today, tomorrow, and in the future. There is an increasing trend in the number of people becoming health/fitness conscious. Most of the shift is attributed to:
– Health risks (e.g. obesity, hypertension, type 2 diabetes) which can cause illness or even death.
– Beauty consciousness where one’s body matters significantly.
Nutrition plays an important role. Food has got a lot of medicinal value as well, making the right food choices can help protect us from common health problems. Also, like taste preferences, every individual has different calorie needs and fitness goals. AI & technology can play a huge role in the following ways:
– Measure the nutritional value of all the food we consume and keep us informed about its health/fitness implications.
– Recommend an appropriate diet for the changing needs of our body.
c) Cooking – Human cooking is often inconsistent. The exact same recipe could taste differently when prepared by different people or differently by the same person on different days. It is important to standardize preparation and cooking like many fast food chains such as McDonald’s. Robots are being developed to automate the process of cooking (checkout Moley). There is an increasing number of startups in this space. In the future, fewer people could be required in the kitchen, maybe even none. AI technology adds a great value if cooking could be assisted with the knowledge of user preferences to make a dish in the way the user expects, this is a way to provide mass personalization.
From a purely combinatorial point of view, though more than ~10^24 recipes are possible (assuming 1000 ingredients and 10 ingredients per recipe, though an actual number of combinations are even more). In reality, only ~10^6 exist. New recipes are constantly being developed by food experts every day; most of it is being done manually and hence exploration is slow. This food alchemy could be made faster & with better intelligence if new recipe discovery is aided with AI.
This is still the tip of the iceberg. We all know for a fact, science & technology changes very fast in every domain. This opens up new and interesting ways in the cross-application of technology to make our lives easier.
Read these articles as a follow up on various points mentioned in this post:
Before you go, tell us how insightful was this article
I’m the CEO & Co-Founder at Spoonshot. Our platform leverages food science and AI to predict consumer needs, F&B trends, and innovation opportunities. We help CPG/FMCG and foodservice companies adopt a data-led, agile, and forward-looking approach to product and menu development.