There’s a great deal of innovation embedded in today’s cutting-edge computer chips, but not much of it is as out-of-the-box as the thinking that’s driving Australian startup Cortical Labs. The company, like so many startups with artificial intelligence in mind, is building computer chips that borrow their neural network inspiration from the biological brain. The difference? Cortical is using actual biological neurons, taken from mice and humans, to make their chips.
“We’re building the first hybrid computer chip which entails implanting biological neurons on silicon chips,” Hon Weng Chong, CEO and co-founder of Cortical Labs, told Digital Trends.
This is done by first extracting neurons in two different ways, either from a mouse embryo or by transforming human skin cells back into stem cells and inducing those to grow into human neurons.
“We then grow those neurons in our laboratory on high density CMOS-based multi-electrode devices that contain 22,000 electrodes on tiny surfaces no larger than 7mm squared,” Chong continued. “These neurons form neural networks which then start to spontaneously fire electrical signals, after a two-week incubation period, that is picked up by our multi-electrode device. The multi-electrode device is also able to provide electrical stimulation.”
The researchers aren’t the first to develop neural networks based on real neurons. Recently, scientists in the U.K., Switzerland, Germany, and Italy fired up a working neural network that allowed biological and silicon-based artificial brain cells to communicate with one another over an internet connection. A California startup called Koniku, meanwhile, is building silicon chips, created using mouse neurons, which are able to sense certain chemicals.
For now, research like Cortical Labs’ is still in a relatively early proof-of-concept stage. According to a recent article in Fortune, Cortical Labs’ current approach has less processing power than a dragonfly brain. That means that, for now, it’s pursuing humbler ambitions than its eventual goal.
“While we’re still in the process of building the hybrid computer chip, right now we’re focused on shaping our neurons’ behavior to play a game of [Atari’s] Pong,” Chong said. “That’s our next big milestone, which will provide a proof-of-concept similar to DeepMind’s demonstration [in 2013] of its A.I. playing Breakout.”
Commercialization is still “a number of years away,” Chong continued. But he’s convinced it could be a game-changer. “When we eventually take our final product to market we believe it will have a wide array of applications across robotics, cloud computing, and computer brain interfaces,” he said. “This does not include industries that we might not have thought about yet because of the novelty of such a computation paradigm.”
AI companies spend enormous amounts of time convincing us their technology is safe and secure. But what happens when those same companies become targets of increasingly sophisticated AI-powered attacks? Anthropic, the company behind Claude, is now facing that question. According to a Washington Post report, cybersecurity startup Cathedral discovered vulnerabilities that let its researchers access sensitive information in Anthropic's internal systems.
The researchers used AI agents to identify potential entry points in less than 12 hours. Interestingly, they didn't need to crack Claude itself. Instead, they exploited weaknesses in third-party business applications that Anthropic uses for everyday operations. Anthropic's security testing program disclosed the findings, which allows outside researchers to identify and report vulnerabilities.
Military working dogs charge into places the rest of us would flee. So, I understand why anyone would want to breed tougher, healthier dogs for the job.
Warhound, a new startup from Colossal Biosciences co-founder Ben Lamm and Anduril co-founder Palmer Luckey, has stepped out of stealth with $200 million to clone and gene-edit working dogs for American troops, police departments, and emergency services (via TechCrunch).
AI can describe a picture, identify objects, and answer questions about what a camera sees. But when a blind user needs a quick answer, getting an entire spoken essay about the scene isn't exactly helpful. Researchers at the University of Maryland (UMD) have found that AI-powered visual assistance can leave blind users doing too much work to get the information they actually need.
The study, titled Say It My Way: Exploring Control in Conversational Visual Question Answering with Blind Users, examined how people interact with Be My AI, an AI-powered feature of the Be My Eyes app. The researchers found that users often received lengthy descriptions that didn't adequately address their immediate goals, forcing them to ask follow-up questions or find ways to steer the AI towards a more useful answer.