Roboticists imagine that, utilizing new AI methods, they will unlock extra succesful robots that may transfer freely by way of unfamiliar environments and deal with challenges they’ve by no means seen earlier than.
However one thing is standing in the way in which: lack of entry to the kinds of knowledge used to coach robots to allow them to work together with the bodily world. It’s far more durable to return by than the information used to coach essentially the most superior AI fashions, and that shortage is without doubt one of the essential issues at present holding progress in robotics again.
Consequently, main corporations and labs are in fierce competitors to search out new and higher methods to collect the information they want. It’s led them down unusual paths, like utilizing robotic arms to flip pancakes for hours on finish. And so they’re working into the identical kinds of privateness, ethics, and copyright points as their counterparts on the earth of AI. Learn the complete story.
—James O’Donnell
My deepfake exhibits how worthwhile our knowledge is within the age of AI
—Melissa Heikkilä
Deepfakes are getting good. Like, actually good. Earlier this month I went to a studio in East London to get myself digitally cloned by the AI video startup Synthesia. They made a hyperrealistic deepfake that seemed and sounded similar to me, with reasonable intonation. The top end result was mind-blowing. It may simply idiot somebody who doesn’t know me properly.
Synthesia has managed to create AI avatars which can be remarkably humanlike after just one yr of tinkering with the most recent era of generative AI. It’s equally thrilling and daunting interested by the place this expertise goes. However they elevate an enormous query: What occurs to our knowledge as soon as we submit it to AI corporations? Learn the complete story.
This story is from The Algorithm, our weekly AI e-newsletter. Join to obtain it in your inbox each Monday.