Cooper Griggs

Precog

20th Century Fox

In an apparent move to feed its smart-hardware ambitions, Google has bought an artificial intelligence startup, DeepMind, for somewhere in the ballpark of $500 million. Considering all of the data Google sifts through, and the fact that it might be getting into robotics, it's not completely absurd that they'd want some software to give a robotic helping hand. (Facebook apparently wanted the company, too, and they've already made moves to wrangle their own sprawling web of information.) But the other part of this story is a little stranger: the deal reportedly came under the condition that Google create an "ethics board" for the project.

What, exactly, does that mean? No idea. It's unclear how the board would be structured, who'd be on it, or when it would be consulted. The London-based DeepMind doesn't seem particularly sinister, either: the company has mostly used its software in fields like e-commerce and gaming. The point is software like this could eventually be used for work in ethical gray areas, and DeepMind might've wanted to get ahead of the issues. 

Which, good. The more decisions we cede to machines, the more we need human oversight of those decisions. A simple "Don't be evil" mantra might not cut it.

XStat

RevMedx

When a soldier is shot on the battlefield, the emergency treatment can seem as brutal as the injury itself. A medic must pack gauze directly into the wound cavity, sometimes as deep as 5 inches into the body, to stop bleeding from an artery. It’s an agonizing process that doesn't always work--if bleeding hasn't stopped after three minutes of applying direct pressure, the medic must pull out all the gauze and start over again. It’s so painful, “you take the guy’s gun away first,” says former U.S. Army Special Operations medic John Steinbaugh.

Even with this emergency treatment, many soldiers still bleed to death; hemorrhage is a leading cause of death on the battlefield. "Gauze bandages just don't work for anything serious," says Steinbaugh, who tended to injured soldiers during more than a dozen deployments to Iraq and Afghanistan. When Steinbaugh retired in April 2012 after a head injury, he joined an Oregon-based startup called RevMedx, a small group of veterans, scientists, and engineers who were working on a better way to stop bleeding.

XStat, before and after

RevMedx

RevMedx recently asked the FDA to approve a pocket-size invention: a modified syringe that injects specially coated sponges into wounds. Called XStat, the device could boost survival and spare injured soldiers from additional pain by plugging wounds faster and more efficiently than gauze. 

The team’s early efforts were inspired by Fix-a-Flat foam for repairing tires. “That’s what we pictured as the perfect solution: something you could spray in, it would expand, and bleeding stops,” says Steinbaugh. “But we found that blood pressure is so high, blood would wash the foam right out.”

So the team tried a new idea: sponges. They bought some ordinary sponges from a hardware store and cut them into 1-centimeter circles, a size and shape they chose on a whim but later would discover were ideal for filling wounds. Then, they injected the bits of sponge into an animal injury. “The bleeding stopped,” says Steinbaugh. “Our eyes lit up. We knew we were onto something.” After seeing early prototypes, the U.S. Army gave the team $5 million to develop a finished product.

But kitchen sponges aren’t exactly safe to inject into the body. The final material would need to be sterile, biocompatible, and fast-expanding. The team settled on a sponge made from wood pulp and coated with chitosan, a blood-clotting, antimicrobial substance that comes from shrimp shells. To ensure that no sponges would be left inside the body accidentally, they added X-shaped markers that make each sponge visible on an x-ray image.

“By the time you put a bandage over the wound, the bleeding has already stopped.”

The sponges work fast: In just 15 seconds, they expand to fill the entire wound cavity, creating enough pressure to stop heavy bleeding. And because the sponges cling to moist surfaces, they aren’t pushed back out of the body by gushing blood. “By the time you even put a bandage over the wound, the bleeding has already stopped,” Steinbaugh says.

Getting the sponges into a wound, however, proved to be tricky. On the battlefield, medics must carry all their gear with them, along with heavy body armor. RevMedx needed a lightweight, compact way to get the sponges deep into an injury. The team worked with Portland-based design firm Ziba to create a 30 millimeter-diameter, polycarbonate syringe that stores with the handle inside to save space. To use the applicator, a medic pulls out the handle, inserts the cylinder into the wound, and then pushes the plunger back down to inject the sponges as close to the artery as possible.

XStat sponges

RevMedx

Three single-use XStat applicators would replace five bulky rolls of gauze in a medic’s kit. RevMedx also designed a smaller version of the applicator, with a diameter of 12 millimeters, for narrower injuries. Each XStat will likely cost about $100, Steinbaugh says, but the price may go down as RevMedx boosts manufacturing.  

If the FDA approves XStat, it will be the first battlefield dressing created specifically for deep, narrow wounds. Gauze, the standard treatment for gunshot and shrapnel injuries, is only approved by the FDA for external use, but “everyone knows that if you get shot, you have to pack gauze into the wound,” says Steinbaugh. When RevMedx submitted its application to the FDA, the U.S. Army attached a cover letter requesting expedited approval. According to Steinbaugh, RevMedx and the military are now in final discussions with the FDA. 

Last summer, RevMedx and Oregon Health and Science University won a seed grant, sponsored by The Bill & Melinda Gates Foundation, to develop a version of XStat to stop postpartum bleeding. In the future, RevMedx hopes to create biodegradable sponges that don’t have to be removed from the body. To cover large injuries, like those caused by land mines, the team is working on an expanding gauze made of the same material as XStat sponges.

“I spent the whole war on terror in the Middle East, so I know what a medic needs when someone has been shot, ” Steinbaugh says. “I’ve treated lots of guys who would have benefitted from this product. That’s what drives me.”

Rose Pastore is an assistant editor at Popular Science. Follow her on Twitter at @RosePastore.

Battling Cancer With A Vaccine

Medi-Mation

Part of what makes cancers so insidious is that they’re not invaders: They’re our own cells turned against us. That means the body usually doesn’t see them as a threat. But over the last few years, teams at several different research institutions have been programming peoples’ immune systems to recognize and destroy cancer. So far, clinical trials of about a hundred terminal leukemia patients have shown some lasting effects. A single treatment has kept two of them cancer-free for three years and counting—after everything they tried had failed. Applying the technique to more cancers requires finding new targets to attack, says Michel Sadelain, an immunologist at Memorial Sloan-Kettering Cancer Center who pioneered the approach. Exploratory clinical trials, including for lung and prostate cancers, are getting under way.

1) Capture T cells (the immune system‘s attack force) from the blood of a patient with B-cell leukemia.

2) Genetically engineer the T cells to train their sights on the CD19 molecule, which sits on the surface of B cells and the cancer cells that arose from them.

3) Inject the patient with the modified T cells, which may then destroy all cells with CD19—both cancerous and not.

4) Bolster the patient’s immune system with treatments of antibodies, since B cells normally make antibodies needed to fight infection.

This article originally appeared in the February 2014 issue of Popular Science.