AI Security Vulnerabilities Surge Tenfold

As the use of artificial intelligence (AI) and large language models (LLMs) expands, security vulnerabilities affecting these services are also increasing rapidly. According to cybersecurity company Kaspersky, the number of vulnerabilities identified in AI and LLM services rose tenfold to 935 in the second quarter of 2026 compared with the fourth quarter of 2025. Of these, 119 were critical security flaws, nearly five times the figure recorded in the final quarter of 2025. The findings were published in Kaspersky’s Quarterly Exploits and Vulnerabilities Report.

The report identified three major categories of security weaknesses in AI-based software: inadequate controls over access to sensitive information and systems, flaws in user authentication, and opportunities to inject malicious instructions into software. Cybercriminals could exploit these weaknesses to gain unauthorized access to systems or carry out harmful activities.

Alexander Kolesnikov, a malware expert at Kaspersky, said, “AI technologies and various tools help make many tasks easier and faster. However, as organizations increasingly adopt AI, new security risks are emerging. System access permissions, user authentication and software vulnerabilities are now major concerns. Staying secure requires more than simply updating software; organizations must monitor their entire technology infrastructure. This will help them identify threats and respond in time.”

Cybercriminals are also exploiting the popularity of AI services in attacks targeting small and medium-sized businesses. According to another Kaspersky report, the company’s security software detected more than 33,300 cyberattacks targeting computers at these businesses during the first four months of 2026.

In these attacks, malicious or unwanted software was presented as software associated with popular AI services. The number of such attacks increased nearly fivefold compared with the same period in 2025.

To address these risks, Kaspersky’s Kaspersky Next security solutions offer continuous monitoring, proactive protection and rapid response capabilities when cyberattacks are detected. According to the company, these features can help organizations using AI protect their technology infrastructure.

//DBTech/SSCPR/MVNO/SMP/DPEO//