Forget Siri, DARPA wants to build a personal assistant that can read your mind
While we are still waiting for Siri to get a little better, the Defense Advanced Research Projects Agency (DARPA) wants to go a step further - they want to build personal assistant machines that can anticipate your needs by reading your mind and body signals.
In a talk at June's DARPA Biology is Technology conference in New York City, DARPA program manager Justin Sanchez explained that the data from our smart watches and trackers actually ends up being pretty meaningless, since these devices can't put data in any sort of context or output a recommended action in response.
"Many of you are just getting things back like 'this is what your heart rate is right now' or 'you took 6,000 steps today,'" he said during a talk. "Who cares about that stuff? What you really want to do is use that information to help you interact with machines in a much deeper way ... today we don't typically aggregate those signals together and do something with it."
With the proliferation of these sensors in smart watches and trackers, Sanchez says it's the right time to develop a smart device that can read these mind and body signals, connect to an external device that makes sense of the information, and then use that information to anticipate what you need and make recommendations.
"We have the pieces," Sanchez told Business Insider. "These sensors are starting to be everywhere. Not only are they in the environment, they're also on our bodies ... we've got the computing power to take the information out of those sensors and we've got the mobile platforms so we have that interface at every step of our everyday lives."
For Sanchez, the possibilities of this technology would be endless.
He points to the Nest Learning Thermostat, which can make changes to the temperature throughout the day based on your past settings, as an example. The Nest Thermostat puts machine learning to practical use, ensuring that after a few days, you may never have to set the thermostat again. Machine learning is a subfield of AI science focused on taking past patterns and making predictions based on those patterns.
Sanchez imagines a "physiological computer" that can read body and mind signals like heart rate and temperature and be able to tell, if you're hot, cold, sleepy, frightened, or bored.
"You could interact with your environment, your architecture," Sanchez said. "Let's say you're having a low point in your day in terms of productivity so what if you had an interface that could say 'how about doing this? maybe this could spark your productivity?'"
Sanchez's talk was called "Brain-Machine Symbiosis," and he suggested that this seamless communication could one day happen directly between the brain and our devices through the use of implantable sensors.
But implants that can detect the brain's electrical signals still have a long way to go.
For one thing, scientists still have yet to design devices that can stay in the brain for a long time without causing damage or losing functionality - the body simply doesn't like them. Surgery to place implants is invasive. Surgeons drill small holes through the skull and "insert long thin electrodes" deep inside the brain," according to a 2012 article in The Scientist.
The implants are often made of stainless steel or other types of metal - useful for conducting electric signals, but problematic for biological purposes.
"If you look at implanted electronics in the brain over the past 10 to 20 years, all suffer from a common problem which is the implant's electronic probes ... create scarring in brain tissue," said Charles Lieber, a chemist from Harvard University who is working on a tiny mesh brain implant.
But that doesn't stop Sanchez. However we get to this "brain-machine symbiosis," he's open to it.
"There are many different futures that can stem from what we and others are doing," he said. "There are a lot of technologies that could potentially get us there, which one is the right one? We can't say. We've just got to try."
Watch Justin Sanchez talk about his ideas at the conference, uploaded to YouTube by DARPAtv: