Plugging Into the Brain: Are You Ready?

Are you ready for computers that can read your mind? Some version of this may not be too far off, according to speakers at this week's Techonomy conference, which began with a discussion of the convergence of man and machine.

Justin Sanchez, who directs the Biological Technologies Office at DARPA, showed several videos that demonstrated technology currently being developed to help soldiers who have been paralyzed. It involves 4mm-by-4mm sensor arrays with 96 sensors that are placed directly on the motor cortex and on the sensory cortex, to control a robotic arm. The signals from the brain are decoded in real time and sent to a robotic arm, and Sanchez played a video of the solider shaking hands with President Obama.

In another video, sensors were placed on the fingertips of the robotic arm, with the signals here transmitted back to the sensory cortex. In this case, the solider could tell which fingers were being pressed, and could even feel as if he were shaking a hand.

This is all very promising for people who have disabilities, but Sanchez, who has been working on neuro-technology for nearly a decade, said these are powerful tools, and we need to consider the ethical, legal, and social implications of their use. "The choices we make about neural technology define who we are today and absolutely will define who we can become in the future," he said.

An opening panel on man and machine began with more prosaic connections, such as voice interface. Rohit Prasad (second from right above), who heads machine learning for Amazon's Alexa, said his focus is on improving daily convenience. Voice is the most natural way of communicating, he said, and talked about the Alexa Prize, a competition for building social bots that can converse with you on Echo for 20 minutes. It will take several years to really get to that point, however, although this year's finalists will be announced at the company's Re:Invent conference.

Tessa Lau (center), the CTO of Savioke, which makes service delivery robots, was described as a "robot whisperer." The company’s first robots do things like deliveries to hotel rooms, and deliberately do not speak. She said that this is just scratching the surface of what the products can do, but added that the company needs more experience of robot-human interaction. To date, she said, the robots have had 200,000 interactions with humans in hotels, and the firm is working to design systems that keep up with the situations that arise in day-to-day interactions.

The most interesting discussion featured Mary Lou Jepsen, who has created a startup called Openwater to work on technology for understanding what people are thinking. Jepsen, who co-founded the One Laptop Per Child project and worked at Google and Facebook, discussed a wearable (something that looks like a ski cap) that would allow communication by human thought alone.

Jepsen said we are seeing a discontinuity with optics that will allow camera chips with LCDs to use infrared light to image blood flow in the brain. This will allow cubic micron resolution, which she said is a million times better than an MRI.

With such technology, Jepsen said, you "can tell what people are thinking," providing a form of telepathy.

There is ongoing research in this area, and she envisions that the first application will be treating brain diseases, such as Alzheimer's, Parkinson's, schizophrenia, and depression. She expects the first products to hit the market in two to three years.

Jepsen acknowledged that both Google X and Facebook turned down her proposals to work on this technology, but said that she thought the physics are inevitable and that something like this will happen within the next 10 years. She said there is a lot of work being done on "reverse engineering the brain," as well as understanding how the firing of neurons equates with thought.

Jepsen discussed the implications of such technology if it could work at a distance, and said "can you imagine just jacking into complete thoughts?" But she said the company is committed to making a product that will only work consensually, at an individual user’s discretion.

But she said such technology could have immense potential to augment intelligence, such as by creating a "rough cut" of a visual idea, music, or writing, just by thinking about it.

Benjamin Bratton of UCSD (far left above) was skeptical, and said he is suspicious of the notion that thoughts are fully formed before language, drawing, etc. Even if we could sense a thought, he said, it would likely be noisy and fuzzy. I'm not certain how close to possible such a technology really is, but it certainly is interesting to think about.