Israel’s cyberattack rate soars as AI tools introduce “hidden” weaknesses
Check Point CPO Nataly Kremer says 10 significant flaws were found in AI applications within a month amid rapid enterprise adoption.
The average organization in Israel experiences approximately 2,000 cyberattacks per week, according to newly presented findings from Israeli cybersecurity company Check Point.
Telecommunications was the most targeted sector, with an average of around 4,000 attacks per week per organization in Israel, compared to a global average of 2,703 attacks per organization. Transportation and Logistics also experienced more than double the global average, with 3,017 weekly attacks per organization, compared with 1,169 globally.
These findings were presented by Check Point CPO Nataly Kremer during Cyber Week at Tel Aviv University.
During her presentation, Kremer also highlighted the major security concerns that come hand-in-hand with organizations’ increased usage of AI tools.
“We are leaping generations ahead in productivity and efficiency, but generations backward in security,” Kremer said.
She reported that within a one-month period, Check Point detected approximately 10 major security flaws in various AI applications, notably including OpenAI Codex.
“Vibe coding tools, which allow any employee to write code quickly and intuitively, create a massive productivity advantage. But alongside this acceleration, new layers of vulnerabilities and weaknesses are exposed within the development environment itself.”
“When AI systems perform actions on behalf of users, they effectively become a new infrastructure layer. Organizations adopting AI without a dedicated security layer expose themselves to training data manipulation, hidden capability injection into models, and information leakage emerging through autonomous agents. Today, protection is required for the models, the training data, the AI supply chain, and the interactions executed on our behalf.”
