Generative AI Continues to Advance
Generative AI rocked the world when it emerged in 2020, prompting a wave of investment from the world’s largest tech companies. Commercially, generative AI has become a valuable B2B product used by many online businesses. It’s a multi-purpose tool that has benefited many online businesses, primarily via advanced chatbots that can better understand context and perform tasks.
While adoption isn’t widespread, future customer-facing businesses are likely to benefit from the technology. Online businesses, whose products or services exist mainly in digital format, stand to gain the most from AI integration. That would be sites offering streaming, user-generated content, e-commerce and iGaming. While popular services like Netflix and Spotify have leveraged the tech, sites that offer video-sharing or slingo bingo games have yet to adopt gen AI for use cases like customer relations. However, as the related technology becomes more advanced and widely available, that may change in the future.
University of Indiana Bloomington Creates Brainoware
In 2022, researchers for the Australian Monash University created working organoids. They were human brain cells grown on a microelectrode dish – dubbed DishBrain by its creators. They trained the organoid to play the 1970 classic, Pong.
— Cortical Labs (@CorticalLabs) October 12, 2022
In late 2023, similar methods were used to combine a brain organoid with a computer, creating the world’s first biohybrid computer called Brainoware. Created by the University of Indiana Bloomington (UIB), Brainoware was grown using stem cells that were connected to a classical computer system.
To test Brainoware, the researchers converted 240 recordings of Japanese speech into electrical pulses. Those speech recordings were split between eight speakers in total, who were instructed to focus on vowel sounds. The organoid responded to each pulse with its own neural activity. Researchers mapped this activity and then used it to train an AI. Parsing through the neural reactions, the AI could then identify speakers based on Brainoware’s reaction to the electrical pulses.
This meant that Brainoware received and understood differences in each electrical pulse, reflecting differences in the speakers’ voices. After starting with a hit-and-miss record near 50%, Brainoware and its corresponding AI improved to 78% accuracy within two days.
Biohybrid Proof of Concept
The Brainoware experiment marks a proof of concept for biohybrid computing. Adherents believe that harnessing brainpower – one of nature’s most powerful and efficient computational mechanisms – could be the key to improving pre-existing systems. At its full potential, some believe that biohybrid computers could surpass modern computers enhanced by developments like generative AI.
However, voice recognition is something that is already being explored by gen AI software created by OpenAI and other industry leaders. Furthermore, gen AI is capable of voice cloning, allowing for accurate repetition of heard speech.
While a useful experiment, Brainoware isn’t feasible for private or public use. It takes a lot of resources to cultivate and maintain neurons, something that isn’t required for gen AI and other advanced software. However, the blending of biological and artificial intelligence raises interesting questions regarding the future of computing and how the legislative landscape will develop around its tech leaders.