Dive Brief:

  • Facebook’s Artificial Intelligence Research (FAIR) group researchers told Forbes they have developed AI software that can determine what ingredients were used to make food in a photograph and provide a recipe.
  • Although it is a feature that many social media users would likely enjoy, Facebook doesn’t currently have any plans to roll out the recipe-generating AI onto its platforms.
  • “We need to have machines that understand the world. Understand not just the visible in the world, but understand that when you have a cake there’s usually sugar in there,” FAIR’s Montreal lab research scientist and project participant Adrianna Romero told Forbes.

Dive Insight:

Accenture predicts that AI will increase economic growth by an average of 1.7% by 2035, and could boost labor productivity by 40% or more. With the anticipated growth of artificial intelligence, companies across the retail sector and beyond are examining how to effectively incorporate the technology into their business models.

Facebook’s experimentation in AI recipe recognition demonstrates that the interest in linking technology with food flows in both directions. On the one hand, social media can benefit from users taking their online interactions into real-world grocery stores. On the other, as online food promotion is driven by images, it’s a natural step for Big Food to encourage the distillation of these popular photos into ingredients — preferably ingredients with their brand attached.

If Facebook is able to perfect the AI behind taking a photo of some red velvet cake, for example, list the ingredients that went into it, and describe the method required to make it, Big Food would be presented with an opportunity to partner with the social media giant to ensure brand-name products are listed as the required ingredients to replicate a delicious looking photo. Especially in a market where center-aisle packaged food sales are suffering, having access to Facebook and Instagram’s wide user base would provide a much-needed surge in visibility for products that customers usually breeze by in person.

Besides building sales, AI that deciphers consumers’ favorite recipes may help build brand loyalty. If customers routinely recognize that their favorite recipes include certain brands, they are liable to reach for the brand name first in the grocery store, knowing that it will likely be used in their kitchen in the near future.

Furthermore, having AI that prompts consumer engagement is an opportunity for companies to gather individual consumer data to help direct future R&D efforts with a better chance of market success, since they’re more likely to be based on real-life culinary interests. This information would not only help internal R&D efforts, but it would also be another tool for large CPG companies with accelerators and incubators. The large companies would have minimal risk in starting their next investments, because they would keep a better pulse on which trends consumers are interested in replicating in their own homes.

At this point, it is unclear how accurate Facebook’s technology is. The social networking giant is also not the only company working on this kind of technology. When MIT researchers announced last July that they’d built a similar system that was trained on a dataset of one million photos and one million recipes, not all the photographs were identified correctly. Still, if Facebook is able to improve upon accuracy, it could stand a chance to change the way consumers search for recipes, as well as buy their groceries.


Sourced from FOODDIVE