Vicarious Develops AI System Capable of Solving CAPTCHA Challenges

Ule Lopez

October 29, 2017

Cybersecurity is one of the biggest and most important matters in the world of technology. Currently many enterprises have suffered the consequences of not listening to warnings by security researchers. As networks have improved and broadened the limits of human interaction, so have increased the risks of attacks and vulnerability exploitations.

Videogames aren’t really alien to this subject either. Gamers are more and more concerned about the safety of their purchases online for digital titles. However, it seems like security will be compromised once again, as there’s a new AI capable of plowing through CAPTCHA challenges with ease.

CAPTCHA systems are those annoying prompts that check to see whether you’re a human or a bot. However, it seems like they have become obsolete with the advent of the new Artificial Intelligence. AI developers have been trying to defeat CAPTCHAs for years, but these systems require a lot of digital horsepower to do so. It might be unbelievable but humans can discern what these AI systems can’t in a matter of seconds.

Enter Vicarious and its Co-founder Dileep George, who trained a computer with neuroscience insights to defeat CAPTCHA challenges. His new system, called the Recursive Cortical Network (RCN), is apparently able to parse the CAPTCHA test more effectively than previous models, and with less training.

This system is incredibly more adaptive and agile than other methods of attempting to conquer CAPTCHA tests. Like the human brain, the new system can apparently learn and generalize using just a few examples.

“CAPTCHAs depend on human vision being better than computer vision.” Mentions Marc Goodman, the author of Future Crimes: Everything Is Connected, Everyone Is Vulnerable and What We Can Do About It. “Unfortunately that’s changing rapidly and computer vision is getting as good as human vision. When that happens, all image-related authentication systems will come under threat as artificial intelligence-based computer vision systems will be able to solve the same exact puzzles that people can.”