Engineers are finding new ways to blend AI with tech breakthroughs. The latest development is a light-manipulating AI chip that is so small it can sit at the end of an optical fiber, yet it is still powerful enough to process information at the speed of light.
Traditional computing relies on electronic circuits, which require significant energy and introduce delays due to the time it takes for data to be processed. However, this new chip operates differently. It uses what engineers call a diffractive neural network.
The engineers built on earlier research from UCLA in 2018, which first demonstrated the potential of diffractive neural networks. Essentially, the diffractive neural network that this light-manipulating AI chip utilizes can drastically speed up how it processes data by harnessing the light itself.
This new version, highlighted in a study published in Nature Photonics, takes that concept further, shrinking it to fit on the tip of an optical fiber, enabling instant data processing with minimal power consumption.
By harnessing light, this chip can process data trillions of times faster than traditional AI hardware, all while consuming only a tiny fraction of the energy. Instead of waiting for a computer to interpret optical signals, the AI chip does the calculations instantly as light travels through it by directly manipulating the light.
This advancement has massive implications for several different fields of technology, including medical imaging, quantum communication, and advanced computing. Using these chips in tiny endoscopic cameras could let medical professionals see into the human body with clearer imaging.
Further, this light-manipulating AI chip could integrate directly with quantum photonic systems—allowing for a massive leap forward in quantum computing. Of course, tech like this always has hurdles to face.
In this instance, the researchers say that manufacturing inconsistencies could cause performance variations. Additionally, because the chip uses a fixed design, it needs to be customized for each individual and specific task. That means scaling up this AI chip that can manipulate light will be difficult.
However, it’s not impossible, though the researchers say overcoming these challenges will take time. Considering some of the other AI advancements we’re seeing lately—like AI-designed computer chips and even how scientists are testing AI’s ability to be sentient—this kind of tech could be closer than ever to becoming mainstream.
