This week - people seem not to be scared by robots that much; a robot learns about itself; tiny robots; China's use of AI worries the "godfather" of deep learning; and more!
Yoshua Bengio, a Canadian computer scientist who helped pioneer the techniques underpinning much of the current excitement around artificial intelligence, is worried about China’s use of AI for surveillance and political control. Bengio said he was concerned about the technology he helped create being used for controlling people’s behavior and influencing their minds. "This is the 1984 Big Brother scenario," he said in an interview. "I think it’s becoming more and more scary."
Here is a 15-minutes long fragment from a podcast where Tom Scott discussed various topics mostly revolving about AI and all the worst-case-AIpocalypse scenarios.
To help algorithms learn some common sense, the Allen Institute for Artificial Intelligence has created an unusual task: pitting machine learning algorithms against humanity in a game of online Pictionary. The program, called Iconary, went online on Tuesday — and you can play it here. I've drawn such a bad tree that the AI didn't know what it was.
Roughly a third of the content published by Bloomberg News uses some form of automated technology. And Bloomberg is not the only one that uses AI to write articles. As the use of artificial intelligence has become a part of the industry’s toolbox, journalism executives say it is not a threat to human employees. Rather, the idea is to allow journalists to spend more time on substantive work.
Google released a white paper to help governments find the best way to regulate AI. Areas where Google invites government rules or guidance include safety certifications for some products with AI inside, like the CE mark used to indicate compliance with safety standards on products in Europe. The white paper offers an example of smart locks that use biometric data, such as face images or thumbprints. The full report (30 pages) can be found here.
This video focuses on tiny robots. Robots so tiny that we can swallow them today but tomorrow millions of them might be inside of us, looking after our bodies and fixing any problems.
World Economic Forum made a survey about globalisation and how people from different part of the world think about problems around climate change, immigration or technology. One of the questions asked was "About how much of what you do in your job do you think could be done today by a machine or robot?". Most people answered that some aspects of their current job are too complicated to be automated. Only 22% of people worldwide think that their job can be fully or almost entirely automated. The full report can be found here (the question about automation is on page 20).
Computerphile visits Dr Henry C. Astley from the Biomimicry Research & Innovation Center at the University of Akron, USA, to learn more about biomimicry and biorobotics and to check out his research on robotic snakes.
To meet its goal of fully driverless cars being tested on public roads by 2021, the UK government announced it’s working on a process to support so-called “advanced trials” of autonomous vehicles — i.e. trials without human safety drivers. The current code of practice allows for automated vehicle trials on any UK road in compliance with UK law — which means test vehicles must include a remote driver. But in the coming years, the government is preparing to drop that requirement.
Researchers at Columbia University reported they have made a step towards a machine self-awareness. They have built a robotic arm that built a model of itself from scratch. After the initial period of learning about itself, its construction and limitations (with no prior knowledge about itself), the robot was able to build a model of itself and perform simple tasks like writing or picking up objects.
That's probably the smallest legged robot ever built. It is too small to have any traditional motors or electronics so its legs are controlled by external magnetic fields acting on tiny cubic magnets embedded in the robot’s hips. In the short interview in the article, Ryan St. Pierre, the creator of this robot, says we can go even smaller.
The U.S. Army has placed a $39 million order for tiny reconnaissance drones, small enough to fit in a soldier’s pocket or palm. The idea behind the drones is that soldiers will be able to send them into the sky of the battlefield in order to get a “lethal edge” during combat.
Leonardo (LEg ON Aerial Robotic DrOne) is a bipedal robot that uses drone-like thrusters not only to help itself keeping the balance but also to increase its agility. The robot can balance itself without the thrusters but adding them into the mix allows Leonardo to be more agile in rough terrain.
Researchers from UC Berkley propose to use newly discovered CasX CRISPR protein instead of widely-used Cas9 protein. CasX seems to be safer for human gene-editing. And I don't think CasX is patented yet (UC Berkley lost the patent for CRISPR-Cas9 to Broad Institute last year).
Chinese researchers claim they’ve built a brain-to-brain interface (BBI) that lets humans guide rats through a complex maze — using only their minds. This isn’t the first time humans have used BBIs to control other animals. Other researchers have used the devices to wag rats’ tails and even control the hands of their fellow humans. But it is the first time a person has used a BBI to complete something as complex as the navigation task set before the Chinese team’s rats — meaning it could put us one step closer to powerful mind-control systems.