This week – mind reading AI; Boening’s heavy drone; drone saves swimmers in Australia; should we replace politicians with robots; ethics of neurotechnologies and AI; and more!
More than a human
Some interesting numbers there. Approximately 76 percent of Americans have heard either nothing at all (40 percent) or not very much (36 percent) about human enhancement technologies. Later we have numbers about acceptance of various types of human enhancements: interventions with respect to vision, joints, cognition and gene editing. Generally, the pattern was the same in each category. The majority was ok with medical applications of human enhancement technologies but when it comes to making healthy people better acceptance was dropping.
Researchers from Alibaba, one of the internet giants in China, have created an AI that outperformed humans at reading with comprehension test. The AI scored 82.44, just past the 82.304 that humans achieved. It’s an interesting news. Some people, however, point out that the AI does not comprehend texts just like humans do.
This article explains why teaching machines to read and, more importantly, to understand what they have read and the able to answer questions about said text is a hard task.
A team from Kyoto University used a deep neural network to read and interpret people’s thoughts. This actually isn’t the first time it’s been done. The difference is that previous methods —and results—were simpler, deconstructing images based on their pixels and basic shapes. The new technique, dubbed “deep image reconstruction,” moves beyond binary pixels, giving researchers the ability to decode images that have multiple layers of colour and structure.
Google Brain team looks back at 2017 and checks what they have done – AutoML, TPUs, Tensorflow and more. Part two, which focuses on our research in the application of machine learning to domains like healthcare, robotics, different fields of science, and creativity, as well as cover our work on fairness and inclusion, is here.
By “leave AI alone”, this article means “don’t regulate AI yet”. The main argument is that the term artificial intelligence is too broad and it is sometimes hard to tell when AI starts. It is a better idea, the argument goes, to focus on specific issues and tailor the solutions accordingly.
Some disappointed with politicians people are asking themselves if an AI would be a better leader than a human.
Companies like Kernel or Neuralink are already working on connecting brains to the internet and augmenting human mind. Not only there are many technical issues to solve but also there are some ethical issues. Not so long ago, a group of researchers called for a new set of guidelines that specifically address concerns that will emerge as groups like Neuralink and other companies around the world explore ways to improve the interface between brains and machines. Their recommendations cover four key areas: privacy and consent; agency and identity; augmentation; and bias.
Boeing debuted a new prototype drone that’s capable of carrying payloads weighing up to 500 pounds (about 227 kg). While such a drone could be mass produced for deliveries, Boeing will instead use the prototype to improve autonomous technology for future products.
We just need a bit more time and resources and who knows, maybe in the future we will be able to put a human brain into a robot the same this digitalised worm brain was put into a Lego robot.
Researchers from UCLA have built a biobot composed of live heart cells, two distinct types of specialized biomaterials for structural support, and flexible electrodes. It’s one of the first examples of bioinspired robotic systems that can imitate nature using both electrical and organic components.
Two teenage boys were rescued by a brand new lifesaving drone in Australia while lifeguards were still training to use the device. Lifesavers sent the drone to drop an inflatable rescue pod, and the pair made their way safely to shore.
There is a loophole in U.S. Department of Agriculture regulations that allows CRISPR-edited plants to fall outside of regulatory purview. While scientists used CRISPR to snip and tweak the plant’s DNA, they did not add any foreign DNA to it, meaning the plants were gene-edited, not genetically “modified,” according to USDA regulations. This new insight into how that regulatory technicality is already making it significantly faster and cheaper to bring new biotech plants to market—shaving years and tens of millions of dollars off production.
Long before CRISPR, there was gene silencing. Unlike CRISPR, which latches onto a gene and snips it out, gene silencing leaves the genome intact. Rather, it targets another molecule—RNA—the “messenger” that shuttles DNA instructions to the cell’s protein-making factory. Now, the gene silencing could make a comeback with a new groundbreaking drug of Huntington’s disease, showing that gene silencing works as a therapeutic strategy.