New Times,
New Thinking.

  1. Science & Tech
30 May 2018updated 24 Jun 2021 12:21pm

How artificial intelligence could personalise food of the future

One day you may be able to hold your smartphone up to a vending machine and order a unique Coca-Cola with a recipe formulated to appeal to your palate.

By Sophie McBain

It might be a triumph of global capitalism that you can order a Coke in London, Tokyo or Ulaanbaatar and, a few minor recipe tweaks aside, it’ll taste pretty much the same. But Jason Cohen, the chief executive of Analytical Flavor Systems, an artificial intelligence platform for food and beverage producers, believes it could soon feel old-fashioned, not to mention unappealing, to consume food and drinks designed for a mass market rather than your specific demographic or – in time – you.

One day you might be able to hold your smartphone up to a vending machine and order a unique Coca-Cola, with a recipe formulated to appeal to your particular palate, and continually tweaked according to your feedback. Cohen’s company uses AI to help food and drink producers appeal to target groups and predict what people might like in the future.

I met Cohen, 27, in a Japanese-style tea shop in downtown Manhattan. He had sandy-coloured hair and a faint blond goatee, and wore a checked shirt and a Buddhist amulet around his neck. Cohen, who is fanatical about tea, guided me through the menu, delivering trivia and the occasional tea-themed joke in his rapid, intense monotone. On his recommendation, I chose a Korean green tea and he ordered the Darjeeling to remind him of his time living on an Indian tea plantation.

When Cohen was still at secondary school, he moved to China to study the language and politics. “Blond hair, blue eyes, bad Chinese doesn’t really endear you to asking about the government,” he said, so instead he began practising his Chinese in the local tea markets. It was the height of the Pu’er bubble, when prices for fermented tea rose to $150 a pound and investors were building special cellars to age it like fine wine. The traders’ passion rubbed off on Cohen, and he later spent a year travelling through the tea-growing regions of China, Tibet, Nepal and India.

After returning to the US to study at Pennsylvania State in 2009, he founded the university’s Tea Research Institute, a group for student tea devotees. The following year he built a computer program that could analyse tea tasting notes in order to identify the type of tea and its country of origin. By 2012, he said, the platform could outperform him and his peers.

Select and enter your email address Your weekly guide to the best writing on ideas, politics, books and culture every Saturday. The best way to sign up for The Saturday Read is via saturdayread.substack.com The New Statesman's quick and essential guide to the news and politics of the day. The best way to sign up for Morning Call is via morningcall.substack.com
Visit our privacy Policy for more information about our services, how Progressive Media Investments may use, process and share your personal data, including information on your rights in respect of your personal data and how you can unsubscribe from future marketing communications.
THANK YOU

Analytical Flavor Systems grew out of this obsession. In 2016 it was accepted to a start-up accelerator run by Techstars and the brewer AB InBev. Cohen is now seeking $2m in funding and hoping to double the company’s workforce to 12.

The team has developed an AI platform called Gastrograph, into which professional tasters input detailed flavour notes and demographic information. Cohen said the program was more effective than the sensory panels traditionally used by food and beverage companies. And it can forecast future flavour trends – such as the current shift in the US from a love of bitter tastes (such as Greek yoghurt and dark chocolate) to a preference for the sour (Icelandic skyr and citrus-flavoured chocolate).

It seems questionable that food forecasting is pure science, rather than partly art or luck – surely our latent desires are not so fixed and formulaic that they can be cracked by a mathematical code? Cohen said he drew inspiration from the streaming company Netflix, which is constantly tinkering with its algorithms to improve the accuracy of its viewer recommendations – and if AI can determine which TV show you’re most likely to binge-watch, surely it can usher in your next ice cream addiction, too.

One disturbing side-effect of AI-generated, personalised foods is that they could perpetuate inequality. Would people accustomed to fatty, sugary, highly processed foods find they are offered less healthy products than consumers who have trained their palates with green juices and raw vegetables?

Cohen acknowledged that this kind of “feedback loop” is a risk, but he believes his technology can have public health benefits, too. Food producers can use it to substitute artificial sweeteners for natural ones without sacrificing flavour, for example. And he believes he is ultimately delivering a social good. “Better tasting products are better for everyone.” 

Content from our partners
Water security: is it a government priority?
Defend, deter, protect: the critical capabilities we rely on
The death - and rebirth - of public sector consultancy

This article appears in the 30 May 2018 issue of the New Statesman, God isn’t dead