The Download: mind-controlled prosthetics, and the price of AI training data
What’s new: When someone loses part of a leg, a prosthetic can make it easier to get around. But most prosthetics are static, cumbersome, and hard to move. A new neural interface connects a bionic limb to nerve endings in the thigh, allowing the limb to be controlled by the brain.
How they did it: First, patients undergo surgery to connect shin muscle, which contracts to make the ankle flex upward, to calf muscle, which counteracts this movement. The prosthetic can also be fitted at this point. In step two, surface electrodes measure nerve activity from the brain to the calf and shin muscles, indicating an intention to move the lower leg. A small computer in the bionic leg decodes those nerve signals and moves the leg accordingly, allowing the patient to move the limb more naturally.
Why it matters: The new device could help people with lower-leg amputations feel as if their prosthesis is part of them, and make walking easier. Read the full story.
—Sarah Ward
AI companies are finally being forced to cough up for training data
The generative AI boom is built on scale. The more training data, the more powerful the model. But many websites and data set owners have started restricting the ability to scrape their websites. Last week three major record labels announced they were suing the AI music companies Suno and Udio over alleged copyright infringement, claiming the firms had made use of copyrighted music in their training data “at an almost unimaginable scale.”
But this moment also sets an interesting precedent for all of generative AI development. Thanks to the scarcity of high-quality data and the immense pressure and demand to build even bigger and better models, we’re in a rare moment where data owners actually have some leverage. The music industry’s lawsuit sends the loudest message yet: High-quality training data is not free. Read the full story.
—Melissa Heikkilä
