The second quarter of 2020 brought a sudden surge in browser-based cryptojacking, with U.S. cybersecurity firm Symantec reporting a 163% increase in detected incidents compared with the previous quarter. To counter this worrying trend, researchers are developing new solutions based on artificial intelligence that could soon become key to stopping criminals from hijacking their victims’ devices to mine cryptocurrency.
Cryptojacking is a term used for unauthorized use of someone’s computer to mine cryptocurrency, usually gaining access by tricking the victim to click on a malicious link, through an infected website, etc.
“After a sharp decline in cryptojacking following the shutdown of browser-based mining script maker CoinHive in March 2019, the second quarter of 2020 saw a resurgence in activity,” Symantec said in its report, demonstrating that the observed spike in activity coincided “with an increase in the value of cryptocurrencies, including Bitcoin and Monero, which are two currencies often mined by browser-based coinminers.”
The number of cryptojacking attacks hit its record level this year in June when it reached a total of 48,697. The surge broke a pattern observed since the beginning of the year when cryptojacking incidents began to gradually decrease, falling from 8,407 attacks in January to 5,403 incidents in May.
Meanwhile, the latest scientific research on AI and how a learning system can be used to detect abusive codes by studying their similarities, brings hope that even more efficient tools against cryptojacking could be in the pipeline.
In a recently published paper in the IEEE Access journal, titled Code Characterization With Graph Convolutions and Capsule Networks, a team of researchers from Los Alamos National Laboratory (LANL) and New York University (NYU) propose the use of an AI-based system to identify illicit crypto mining by comparing its code to its legitimate counterpart.
“Our deep learning artificial intelligence model is designed to detect the abusive use of supercomputers specifically for the purpose of cryptocurrency mining,” said Gopinath Chennupati, a researcher at LANL and co-author of the paper.
The researchers claim that, as all programs can be represented by graphs that comprise nodes linked by lines, loops, or jumps, their AI system could be used to compare “the contours in a program’s flow-control graph to a catalog of graphs for programs that are allowed to run on a given computer.”
“Based on recent computer break-ins in Europe and elsewhere, this type of software watchdog will soon be crucial to prevent cryptocurrency miners from hacking into high-performance computing facilities and stealing precious computing resources,” according to Chennupati.