This week - delivery robot disrespects crime scene in LA; Google is making LaMDA available; how AI is changing protein design; and more!
More Than A Human
Replacing human body parts with bionics is a staple of science fiction. But it might soon change. The progress in soft robotics, material science and embodied intelligence leads one researcher to predict that within 10 years, the technology will advance enough to "restore functionalities of the human body through bionic parts".
LaMDA, Google's conversational AI that led one of their engineers to claim it is sentient, is going to be available for people to play with and provide feedback. I'm linking this article by Alberto Romeo who does a good job at recapping the controversy around LaMDA and AI's limitations.
Researchers from the University of Oxford and Google DeepMind came to the conclusion that an existential catastrophe coming from a super-advanced AI is now "not just possible, but likely". Their paper proposes that at some point in the future, an advanced AI overseeing some important function could be incentivized to come up with cheating strategies to get its reward in ways that harm humanity.
If you want to build a good mind-reading machine, your best approach right now would be to insert an implant into a brain to listen to brain signals. Researchers from Meta looked at the possibility of replacing the invasive procedure of putting sensors into a brain with an AI that listens to brainwaves via a non-invasive brain-computer interface. Initial results are promising.
AlphaFold and RoseTTAFold completely changed how scientists work with protein design. Things that previously took months can be done in minutes or seconds. This article published in Nature describes the ongoing revolution in protein design and the efforts of many scientists to improve the workflow and ultimately make the proteins dreamed up by an AI in a lab.
Here is yet another example that the real world is much more messier and unpredictable than people think. I don't think the designers of this food delivery robot thought their creation will ever encounter a crime scene and taught it what to do in this situation.
Laughter comes in many forms, from a polite chuckle to a contagious howl of mirth. Scientists are now developing an AI system that aims to recreate these nuances of humour by laughing in the right way at the right time. The team behind the laughing robot, which is called Erica, say that the system could improve natural conversations between people and AI systems.
Researchers used a series of synthetic genetic circuits to grow plants with modified root structures. Their work is the first step in designing crops that are better able to collect water and nutrients from the soil and provides a framework for designing, testing, and improving synthetic genetic circuits for other applications in plants.