AI automatically evolves to circumvent censorship
The new tool already proved to be successful in China, India, KazakhstanEuropost
Internet censorship by authoritarian governments prohibits free and open access to information for millions of people around the world. Attempts to evade such censorship have turned into a continually escalating race to keep up with ever-changing, increasingly sophisticated internet censorship.
Censoring regimes have had the advantage in that race, because researchers must manually look for ways to circumvent censorship, a process that takes considerable time.
But new work led by University of Maryland computer scientists could shift the balance of the censorship race. The researchers developed a tool called Geneva (short for Genetic Evasion) which automatically learns how to circumvent censorship. Tested in China, India and Kazakhstan, it successfully evaded censorship in all of these countries by exploiting gaps in censors' logic and finding bugs that the researchers say would have been virtually impossible for humans to find manually.
“With Geneva, we are, for the first time, at a major advantage in the censorship arms race,” said Dave Levin, an assistant professor of computer science at UMD and senior author of the paper. “Geneva represents the first step towards a whole new arms race in which AI 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,” he added.
Known as a genetic algorithm, Geneva is a biologically inspired type of AI that Levin and his team developed to work in the background as a user browses the web from a standard internet browser. Like biological systems, Geneva forms sets of instructions from genetic building blocks. But rather than using DNA as building blocks, it uses small pieces of code. Individually, the bits of code do very little, but when composed into instructions, they can perform sophisticated evasion strategies for breaking up, arranging or sending data packets. Additionally, the system mutates and crossbreeds its strategies by randomly removing or adding new instructions, or combining successful ones before testing the strategy again.
As Levin states, this unique approach allows researchers to not only identify how a censorship strategy works, and then devise ways to evade it, but also shows them what censorship strategies are being used, by seeing how Geneva defeated them. Moreover, when Geneva is running on a computer that is sending out web requests through a censor, it modifies how data is broken up and sent, so that the censor does not recognise forbidden content or is unable to censor the connection.
The researchers now plan to release their data and code in the hopes that it will provide open access to information in countries where the internet is restricted. The team is exploring in particular the possibility of deploying Geneva on servers rather than on client computers. That would mean websites such as Wikipedia or the BBC could be available to anyone inside countries that currently block them, without requiring the users to configure anything on their computer.