CAPTCHA is an acronym for Completely Automated Public Turing test to tell Computers and Humans Apart. The term was coined in 2003, when the use of automated bots was becoming commonplace, and it refers to those annoying squiggly distorted letters that you have to type in when creating an online account. Although some companies have found ways around them, CAPTCHAs are still ubiquitous online.
Researchers at the AI company Vicarious may have just made CAPTCHAs obsolete, however, by creating a machine-learning algorithm that mimics the human brain. To simulate the human capacity for what is often described as “common sense,” the scientists built a computer vision model dubbed the Recursive Cortical Network.
“For common sense to be effective it needs to be amenable to answer a variety of hypotheticals — a faculty that we call imagination,” they noted in a post at their blog.
The ability to decipher CAPTCHAs has become something of a benchmark for artificial intelligence research. The new Vicarious model, published in the journal Science, cracks the fundamentals of the CATCHPA code by parsing the text using techniques that are derived from human reasoning. We can easily recognize the letter A for example, even if it’s partly obscured or turned upside down.
As Dileep George, the co-founder of the company explained to NPR, the RCN takes far less training and repetition to learn to recognize characters by building its own version of a neural network. “So if you expose it to As and Bs and different characters, it will build its own internal model of what those characters are supposed to look like,” he said. “So it would say, these are the contours of the letter, this is the interior of the letter, this is the background, etc.”
These various features get put into groups, creating a hierarchal “tree” of related features. After several passes, the data is given a score for evaluation. CAPTCHAs can be identified with a high degree of accuracy. The RCN was able to crack the BotDetect system with 57 percent accuracy with far less training than conventional “deep learning” algorithms, which rely more on brute force and require tens of thousands of images before they can understand CAPTCHAs with any degree of accuracy.
Solving CATCHPAs is not the goal of the research, but it provides insight into how our brains work and how computers can replicate it, NYU’s Brenden Lake told Axios. “It’s an application that not everybody needs,” he said. “Whereas object recognition is something that our minds do every second of every day.”
“Biology has put a scaffolding in our brain that is suitable for working with this world. It makes the brain learn quickly,” George said. “So we copy those insights from nature and put it in our model. Similar things can be done in neural networks.”
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.