The courage to break convention

Features - Cover Story

MIT’s OpenAG initiative is challenging traditional limits to information sharing with the launch of a collaborative database of plant “recipes.”

March 22, 2016

A few years ago, Caleb Harper was curious to know what would happen if data scientists and engineers channeled their talents into elevating controlled environment agriculture (CEA) crop quality — and what he found was that it would look entirely different than what farmers were accustomed to.

Harper, a former architect and engineer for hospital data centers, is now the principal investigator and director of Massachusetts Institute of Technology’s (MIT) OpenAG initiative, which will host a crowdsourced collection of specific crop production “recipes,” or combinations of climate conditions and crop inputs into an open-source data platform. These recipes will be tested in three different scales: a personal “food computer,” labs the size of a shipping container called “food servers,” and a “food data center,” which will be a climate-controlled warehouse in Boston.

The end goal is to give farmers worldwide, large and small, the opportunity to discover better ways of perfecting their varieties — while contributing discoveries of their own.

During Harper’s latest TED Talks presentation filmed in Geneva, Switzerland in December 2015, he says, “The future of food is about networking the next 1 billion farmers, and empowering them with the platform to ask and answer the question, ‘What if?’”

Power in numbers

An overview of OpenAG's food server, located inside MIT's media lab.
Photo courtesy of OpenAG

The idea of OpenAG stemmed from Harper’s trip to Japan after the nuclear disaster at Fukushima in 2011. Moved by the worries that agricultural land may no longer be safe for use, he set out to find a way for his data-driven background to be useful in sustainable food production.

From there, he and a few MIT researchers created a small-scale vertical farm to grow just enough food to feed MIT’s media lab — about 300 people. The farm itself uses sensors to monitor 30 different variables within an edible plant, such as humidity, CO2, light levels, EC, temperature, etc., and stores them in the so-called “recipe” platform. The database will provide growers with detailed metrics of the measurements it took to make a tomato or lettuce variety taste so sweet (or not so sweet), for example, and replicate the exact process if they so desire.

When it was time to harvest, those in the media lab picked the heads of hydroponic lettuce and ripped their roots off before eating them. The process piqued the interest of mechanical engineers who also wanted to collaborate with the project.

A front-facing view of the food server, which is used to research new and contributed "plant recipes."
Photo courtesy of OpenAG

Others without farming backgrounds, including tech students, became curious about growing, like a 17-year-old back-end coder, who found he could control the plant’s health with the code he was building. “Do you know what happens when his code breaks now? He kills things,” Harper jokes in a Wired UK presentation published in October 2015. “His code has now become so important and such a part of the biologic process that he sees how things can be related,” he adds.

Grow your own recipe

"We are going to learn about a complexity that we can only guess at now," Harper said in his most recent Ted Talk.
Photo courtesy of OpenAG

The OpenAG team, now comprised of about 20 top-tier engineers, scientists and architects — with advisers like vertical farming pioneer Dickson Despommier, Howard Shapiro, Mars’ global director for plant science, and Toyoki Kozai, president of the Japanese Plant Factory Association — is making its open source mission more tangible with the development of at-home farms, or “food computers” as MIT calls them.

These buildable, tabletop-sized, enclosed units can be made anywhere in the world from tools and materials that can be purchased at a hardware store, so that others may experiment with current plant recipes as well as create new, unique ones. LED lights can be installed at the top of the unit, and users will be able to track their progress with the aforementioned sensors.

As the project stands now, a person can build this computer if he or she is highly skilled and has access to a laser cutter. However, OpenAG is working on Version 2 of the food computer, which Harper says will be cheaper, less challenging to build and easier to maintain.

So far, about 30 food computers exist globally, and users are already contributing quite a bit of data into the system, according to Harper. Ten computers have also gone to elementary schools in the Boston area for additional experimentation.

It is my genuine feeling, that what I'm doing will be good for all of us. I think it will help us scale. I think it will help us focus, to have some baseline technologies that everybody is using.” — Caleb Harper, director of OpenAG

“Thirty little farms in aggregate tells you an awful lot about what needs change, what’s working, what’s failing,” he said in a phone interview with Produce Grower. And the best part? “We can run analytics on it,” Harper adds, so that results may be tracked and studied for further improvement.

Harper is currently working with Mozilla, the open standard internet browser company, to draft a license agreement for OpenAG’s database users so that they may agree to keep what they find as open source, and contribute ideas and recipes of their own.

Scale up

Photo courtesy of OpenAG

Larger concepts that build upon the food computer, such as what OpenAG has dubbed the “food server,” are also in the works. This medium-sized scale (40 feet by 8 feet by 10 feet) is like a mini-version of the upcoming “food data center.” The food servers are meant to “appeal to interdisciplinary researchers and small-scale cafeterias, restaurants and boutique operators,” according to the OpenAG website.

These CEA models, be they the food computers or food servers, are already located in India, Ghana, Australia, Hong Kong, Italy and Mexico, and a handful of locations across the United States, including Detroit, San Francisco, Cambridge and Minneapolis. And the warehouse-turned food data center is slated to be complete in about a year and a half, Harper says.

Harper and his team area also working together with a non-food industry partner, the global textile manufacturer WelSpun, to experiment with this type of crop monitoring in “food server” sizes in India. The operations are testing cotton grown in controlled environments to “produce a superior-quality expression, or in this case, a superior quality fiber, longer cotton, longer staple lengths, more inherent pigments, more uniform fiber [and] more tensile lengths,” Harper says.

USDA has tested the data for crops tested so far, and while results have not yet been published, Harper claims that “the cotton we produced was as good or better than the cotton in the market that is spinable cotton.” And plans are in the works for an additional site where WelSpun can do some of their own testing.

What open source could mean

The instructions to build a personal food computer are located on OpenAG's website,
Photo courtesy of OpenAG

So far, the project has received sponsorship and/or philanthropic backing from several major players — such as Target, General Electric, Mozilla, Google, Unilever, National Geographic and General Mills, but Harper has also received pushback and criticisms of the OpenAG initiative from some CEA farmers, and others in the industry.

“There are companies with very vested interests [in] keeping the data of a climate recipe, or a light recipe, or a fertilizer recipe proprietary because they took a long time to try to find it. There are companies exploring patents on those recipes now,” he says. “We’ve sold the idea of proprietary systems. We have to keep with that to keep getting funding. And that is the limiting factor. That doesn't mean it’s right, but I think it’s limiting because we can’t learn from our mistakes.”

And while he doesn’t necessarily believe the sole way to solve the world’s food crisis is CEA, Harper believes the industry as a whole won’t be able to improve if it continues to think in proprietary terms of the same basic systems and technology.

In the world of CEA, Harper says, “There are special tweaks here and there, and there are great insights on one [particular] thing, but there are not necessarily revolutionary systems that are so completely different from each other, and good, and functioning, and making everybody money, and growing as we might be led to believe right now.”

In the meantime, Harper and his team are focusing their efforts to get growers on the same page.

“It is my genuine feeling that what I’m doing will be good for all of us,” he says. “I think it will help us scale. I think it will help us focus, to have some baseline technologies that everybody is using.”