Artificial intelligence to detect and avoid internet censorship

Oppressive governments around the world monitor and control the Internet materials that are sent to and received by their citizens.

Trying to avoid such censorship is a slow and difficult process. Often, researchers need to find ways to bypass it.

But computer scientists at the University of Maryland have developed an artificial intelligence program called Geneva, which understands how censorship technology is used and adapts in real time to bypass it.

“I think Geneva is changing this game because it finally allows sites to react very, very quickly to changes in censorship,” explains Dave Levin at the University of Maryland.

The information on the Internet is shared into a host of data from the sender’s computer that is reconnected by the receiving computer. Repressive regimes monitor data packets and block requests that contain certain words (such as Tiananmen Square in China) or banned sites such as Wikipedia in many countries.

Geneva reacts automatically to circumvent this kind of censorship in several ways.

“One way, it can make the censor think, ‘it’s not something to be censored, it’s appropriate information, or it tricks the censor into thinking he broke the connection, cut off communication, when in fact he did not really do it.’ explains Dave Levin.

For example….

“About a year ago, China introduced a completely new censorship system to try to interrupt certain types of secure communications. “It only took about an hour for our program to find ways to bypass it,” said Kevin Bock of the University of Maryland.

Furthermore, users do not need to install the program. It can run in the background of an application.

“This was very important for us, because it means that people can avoid censorship without having to take any risks to install the program,” says Dave Levin / VOA /

top channel