Over the past five weeks, a limited number of hackers breached more than 600 firewalls in dozens of countries with the help of widely available artificial intelligence tools, according to Amazon.com Inc.’s security research.
According to a company report, small groups of hackers – or possibly just one person – used commercial generative AI services to exploit weak security measures, such as simple sign-in credentials or single-factor authentication. The techniques allow intruders to compromise firewalls on a scale that would otherwise require a larger and more skilled team. Russian-speaking hackers leveraged their access to security devices spread across 55 countries to move more into some victims’ networks, orchestrating ransomware attacks, the report said.
The widespread breach, which Amazon says was financially motivated, is the latest example of hackers using AI to simplify and speed up cyberattacks.
“It’s like an AI-powered assembly line for cybercrime, helping to produce low-skilled workers at scale,” CJ Moses, who leads security engineering and operations at Amazon, said in the report. It does not identify the AI tools used by the hackers, nor does it name the victims.
Rather than opportunistically targeting certain industries, researchers believe hackers breached firewalls with weak protections, according to the report. Compromising instruments were spread across South Asia, Latin America, the Caribbean, West Africa, Northern Europe and Southeast Asia.
When hackers faced tighter security, they simply moved on to other targets, Moses said. And once inside the network, they “were largely unsuccessful when attempting to exploit anything beyond the most straightforward, automated attack paths,” the report said.
Last year, a hacker exploited technology from Anthropic PBC as part of a broader cybercrime scheme that affected at least 17 organizations, Anthropic said, marking an “unprecedented” example of attackers weaponizing a commercial artificial intelligence tool on a large scale.
Amazon expects more to come.
“Organizations must anticipate that AI-enhanced threat activity from both skilled and unskilled adversaries will continue to increase in volume,” said Moses.
Topics
InsurTech Cyber Artificial Intelligence
I am interested. Ai?
Get automatic alerts for this topic.
