The first poop database in the world needs your help

Artificial intelligence can soon decode your poop. That is the ultimate goal of a campaign to collect 100,000 fecal photos to build what, according to developers, is the first database of poop images in the world. The campaign, launched by microbial health company Seed, challenges you to give "shit" to science by uploading photos of your feces so that scientists can use it to train an AI platform launched from MIT .


In this case, offering your stool to science may help around 1 in 5 people in the US who have chronic bowel disorders such as irritable bowel syndrome, developers behind the database say. They hope it will give doctors an additional diagnostic tool and enable patients to better manage their health.

"We don't always think about stool as a daily data – I put air quotes around" dump "- but actually, as a direct output of our gut health," Ara Katz, co-founder and co-CEO of Seed Health, says The edge.

This is how civil scientists can contribute to the cause. To participate, go to on your phone (because it's weird to bring your laptop to the bathroom and you can't submit a photo on the page unless you use your phone). Click on the large purple button with the text & # 39; # GIVEaSHIT & # 39 ;. You will be asked to enter your e-mail address and whether you have a morning, afternoon or evening poo. If you have already dropped a deuce, you can take your photo or upload it, or you can request an email reminder to send you to the time you specified. After submitting your stool for posterity, the image will be separated from the metadata (your email address and other potentially identifying information) so that your donation can remain anonymous and HIPAA compatible.

A team of doctors will diligently view every received image. (Yes, that's a real job for seven gastroenterologists who take notes about what they see in the photos.) Poop can be identified in seven categories identified along the Bristol scale that you and your doctor can tell if you are clogged, lack fiber, have a serious case of the runs, or somewhere in between. The insights of the doctors in your poop help train artificial intelligence models to understand the same things that the doctors see in the picture. Similar training systems are used to teach self-driving cars how to identify a tree or cat on the road, according to David Hachuel, a co-founder of the startup Auggi, who is building the platform.

Because there is no existing database of real poop, the Hachuel team started developing the app using Play-Doh that they had designed to simulate poop around the Bristol scale. "We have spent countless hours making different Play-Doh models," Hachuel says The edge. "We actually printed a toilet 3D to see how that would happen in real life." That was nice, he says, but time-consuming. To understand the real thing, they first turned to Reddit and other sites to crowdsourcing photos of droppings that people had uploaded to the discussion board. Both Hachuel and Katz tell The edge that you would be surprised to see how many people already share and discuss their poop online.

Fun and aside, Hachuel says that the aim of the campaign is to ultimately make the database an open-source tool for academic researchers. In the meantime, the hope is to enable patients to regain control of their own gut health. He says they are building technology that helps people with chronic bowel disease to better understand the relationship between their lifestyle and their symptoms. "They struggle every day with making decisions about what to eat, how much exercise they have to do to keep their symptoms at bay," he says. "And so it is very important to build this database and to develop these simple monitoring tools to enable those patients to do it essentially at home."