In the quest for more energy-efficient artificial intelligence, some scientists are delving into the realm of living computers. These innovative systems are designed to use biological matter instead of traditional silicon-based hardware. This approach, known as biocomputing, involves the use of synthetic biology to create computer architecture using miniature clusters of lab-grown cells called organoids.
One company at the forefront of biocomputing is FinalSpark, which has introduced the Neuroplatform. This platform is powered by human-brain organoids and can be rented by scientists over the Internet for $500 a month. The goal of FinalSpark is to create artificial intelligence that consumes 100,000 times less energy than current state-of-the-art AI models. The Neuroplatform consists of processing units hosting spherical brain organoids that are connected to electrodes for stimulation and network connectivity.
The organoids are exposed to dopamine to mimic the brain’s reward system, which helps train the neurons to form new connections and pathways. This training could potentially enable organoids to function as processing units similar to CPUs and GPUs in traditional computers. While the technology is still in its early stages, researchers can observe the behavior of the organoids 24 hours a day to study and understand how they function.
FinalSpark has partnered with research teams from various universities around the world to explore the possibilities of biocomputing. These teams are investigating different aspects of organoid activity and how they can be integrated into AI learning models. While there are challenges, such as the limited lifespan of organoids, researchers are optimistic about the potential of biocomputing to revolutionize computing.
Apart from brain organoids, other forms of biocomputing are being explored. Researchers are studying the use of modified living cells to create computing systems that can replicate memory, logic gates, and decision-making processes found in conventional computers. This approach holds promise for applications like bioremediation, where conventional computers fall short.
While there are ethical considerations surrounding the use of human neurons for nonmedical purposes, researchers are actively engaging in discussions to address these concerns. The potential of biocomputing, whether through brain organoids or other biological systems like fungi, opens up new possibilities for the future of computing. By leveraging the unique capabilities of living organisms, scientists are paving the way for a new era of energy-efficient and sustainable artificial intelligence.
