Chinese scientists have developed brain-implantable processors that serve artificial intelligence with minimal power consumption, marking a significant advancement in chip efficiency. Engineers from the Chinese Academy of Electronics and Information Technology assert the creation of the most energy-efficient processors to date, aimed at supporting artificial intelligence applications.

"These processors are designed to be embedded in as many devices as possible. However, their substantial computational capacity often requires significant energy, limiting their usability," notes Interesting Engineering. During the International Solid-State Circuits Conference (ISSCC), an annual gathering of the world's leading researchers, engineers, and professionals in chip technology, the team unveiled two groundbreaking chips.

The first chip is engineered for integration into smart devices to enable offline voice control. According to the engineers, this chip can recognize the speaker's voice even in noisy environments, such as when the television is on, music is playing, or others are talking. It consumes less than two microjoules of energy for voice recognition, achieving an accuracy of 95% in quiet and 90% in noisy settings, as per team leader Zhou Chun.

The second chip, also presented at ISSCC, is designed to detect epileptic seizures. This chip, intended for wearable devices, uses electroencephalography (EEG) to identify signals of epilepsy and alerts the patient to seek medical assistance or apply a predetermined therapy. This chip is touted as having the most energy-efficient design, with an average recognition energy consumption of only 0.07 microjoules, reported by South China Morning Post.

Additionally, this chip can serve as a brain-computer interface. In a demonstration, a subject was able to control the movement of a robot with their thoughts, showcasing the chip's potential applications beyond health monitoring.