Researchers have built an artificial intelligence (AI) tool which can automatically learn and adapt itself to circumvent censorship on the internet, an advance that may open up blocked online content for millions of people living in India and China.
The researchers, including those from the University of Maryland (UMD) in the US, tested the technology in China, India and Kazakhstan, and found dozens of ways to overcome censorship by exploiting gaps in logic used by censors, and finding bugs that are difficult for humans to find manually.
The researchers said they plan to introduce the tool called Geneva -- short for Genetic Evasion -- during a peer-reviewed talk at the Association for Computing Machinery's 26th Conference on Computer and Communications Security in the UK on Thursday.
"With Geneva, we are, for the first time, at a major advantage in the censorship arms race," said Dave Levin, senior author of the paper from UMD.
"Geneva represents the first step toward a whole new arms race in which artificial intelligence systems of censors and evaders compete with one another. Ultimately, winning this race means bringing free speech and open communication to millions of users around the world who currently don't have them," Levin said.
The researchers said one of the common forms of internet censorship, particularly used by authoritarian regimes, monitored the pieces of information sent during an internet search called data packets.
They said censors block requests that either contain flagged keywords -- such as "Tiananmen Square" in China -- or prohibited domain names like "Wikipedia" in many countries.
When a computer running Geneva sends out web requests through a censor, the AI tool modifies how the data is broken up and sent, in a way that the censor does not recognize as forbidden content.
Due to this, the researchers said, the censor is unable to block the connection.
They said Geneva is a biologically inspired type of artificial intelligence.
With each generation, Geneva keeps the instructions that work best at evading censorship, and kicks out the rest, evolving to be better each time like a biological system, they explained.
Geneva crossbreeds and mutates its strategies by removing instructions randomly, and adding new instructions, the researchers said.
They added that the technology also combines successful instructions and tests its strategy in recursion.
Through this evolutionary process, the researchers said, Geneva is able to identify multiple evasion strategies from censors very quickly.
According to the researchers, the AI tool forms sets of instructions from building blocks that are similar to a genetic code.
They said, instead of using DNA as building blocks, Geneva used small pieces of computer code.
The bits of code do very little by themselves, but when composed into instructions, they can perform sophisticated evasion strategies for breaking up, arranging or sending data packets, the researchers said.
"This completely inverts how researchers typically approach the problem of censorship," said Levin.
"Ordinarily we identify how a censorship strategy works and then devise strategies to evade it. But now we let Geneva figure out how to evade the censor, and then we learn what censorship strategies are being used by seeing how Geneva defeated them," he said.
The research team ran Geneva on a computer in China with an unmodified Google Chrome browser installed, and found that they were able to browse free of keyword censorship.
They could also use the tool to evade censorship in India, where they noted that "forbidden URLs" are blocked.
In Kazakhstan, where there is prevalent eavesdropping on certain social media sites, the researchers said they could use Geneva to successfully circumvent censorship.
"Currently, the evade-detect cycle requires extensive manual measurement, reverse engineering and creativity to develop new means of censorship evasion," said Kevin Bock, lead author of the study at UMD.
"With this research, Geneva represents an important first step in automating censorship evasion," Bock said.
Disclaimer: No Business Standard Journalist was involved in creation of this content
