At the Paris Air Show this summer, Airbus trumpeted a $170,000 home-automation system called Swirl, which the aerospace manufacturer believes will be installed in every new Airbus on Earth by 2022. A crew of more than 20 pilots, scientists, and students have come up with the project. The kitchen, apartment, and bedroom systems are overseen by AI neural networks, and sensors and computer vision image processors make up the rest of the package.
Well, Airbus actually doesn’t say so, but rather repeats that the system is, on a whole level, a “gamification” of the satellite industry’s standard satellite constellation of satellites up there in space. And you read that right. Spaceflight is already, in part, gaming. On our planet, it’s arguably illegal to use the Internet without having access to the proper chips or browsers. It’s common for people to not just use the network for fun, but for business; they are what you might call “entrepreneurs on steroids”.
The Fringe Social Science Of Interest
Another example of the latest thinking in artificial intelligence came from MIT, where Ahmed Mashael and Erica Landau presented their work on “sentient classrooms.” The pair published their paper in Technology Review earlier this month, and it’s featured in this week’s issue of IEEE Spectrum. The broader issue of AI, along with the projects in robotics, and other areas, is that we need to find a way to harness and translate AI skills into social good. (See the related blog posts inside.)
At MIT, Mashael and Landau found that without spending hundreds of millions of dollars–besides which, “artificial intelligence doesn’t work with robots,” Mashael tells me–teachers can create “androids that can learn something, manage their educational environment, and manage and extend knowledge of the world.”
The pair analyzed video footage from the 2014 Geneva Motor Show, where an upcoming all-electric car was on display. (There, a car company displays “algorithms, algorithms, and more algorithms,” Landau says.) They found plenty of examples of robots interacting with humans, but they didn’t find much patterning. What the scientists did found, says Mashael, is “exotic methods of integrating intelligence into natural domains.”
For example, he points to the 2011 Car2Go demonstration car, which started to drive itself following autonomous driving techniques. And there are examples of robots helping disabled people or people afflicted with epilepsy. Many of these helpfully act as caregivers, by picking up medicines, cleaning and wiping clothes, and more. They are examples of A.I. doing what humans do on purpose, rather than inadvertently.
To see more of Mashael and Landau’s work, check out their MIT Sloan Educator’s page, and coverages at USA Today, and MIT Tech Review.