Chennai: A new class of attacks in the digital world known as ‘DeepAttacks’ using AI-generated content will emerge to evade AI security controls, according to a report.
Released by digital security products major Avast, the annual Threat Landscape Report, states that last year, the Avast Threat Labs team observed many examples where researchers used adversarial AI algorithms to fool humans. Examples include the fake Obama video created by Buzzfeed where President Obama is seen delivering fake sentences, in a convincing fashion.
The report which talks about the security trends facing consumers in 2019, looks into topics such as new form of attacks, IoT threats, router attacks, mobile threaats, among others.
The dawn of a adversary
The report states that Avast foresees the emergence of a class of attacks known as ‘DeepAttacks’, which use AI-generated content to evade AI security controls.
“Avast has seen examples of adversarial AI deliberately confounding the smartest object detection algorithms, such as fooling an algorithm into thinking that a stop sign was a 45-mph speed limit sign,” the report said and added that DeepAttacks are expected to be deployed more commonly in an attempt to evade both human detection and smart defences.
Sophisticated threats
The trend toward smart devices will be so pronounced in the coming years that it will become difficult to buy appliances or home electronics that are not connected to the internet, the report said.
“Security is often an afterthought in the manufacturing of electronic devices. While the big-name smart devices often do come with embedded security options, some producers skimp on security either to keep costs low for consumers or because they are not experts in security,” it stated.
It stated that IoT malware can be expected to evolve and become more sophisticated and dangerous, similar to how PC and mobile malware developed.
Advanced attacks
Routers have proven to be a simple and fertile target for a growing wave of attacks. Not only have we seen an increase in router-based malware in 2018, but also changes in the characteristics of those attacks.
“It can be expect to see the increased hijacking of routers used to steal banking credentials. This new element could ask mobile users to install a new banking app, later capturing authentication messages,” the report said.
Routers will continue to be used as targets of an attack, not just to run malicious scripts or spy on users, but also as an intermediate link in chain attacks, it added.
Evolution of mobile onslaught
“In 2019, well known tactics such as advertising, phishing and fake apps will continue to dominate the mobile threat landscape. Fake apps are the zombies in mobile security. They will continue to persist as a trend in 2019,” the report said.
In 2018, the return of banking Trojans was also particularly pronounced on the mobile side, growing 150 per cent year-on-year, from three per cent to over seven per cent of all detections worldwide, the report said and declared that cybercriminals are finding banking to be a more reliable way to make money than cryptomining.
| Expert speak |
| “This year, we celebrated the 30th anniversary of the World Wide Web. Fast forward thirty years and the threat landscape is exponentially more complex, and the available attack surface is growing faster than it has at any other point in the history of technology.” said, president of Consumer, Avast, Ondrej Vlcek.
“PC viruses, while still a global threat, have been joined by a multitude of malware categories that deliver more attacks. People are acquiring more and varied types of connected devices, meaning every aspect of our lives could be compromised by an attack. Looking ahead to 2019, these trends point to a magnification of threats through these expanding threat surfaces.” |