As corporations expand adoption of artificial intelligence (AI), information security spending is rising alongside it. "Shadow AI," in which employees use external AI services outside a company's control, is also emerging as a new security risk.
Kaspersky said on the 17th that, in a global survey of 1,800 IT and cybersecurity decision-makers and experts, 41% of responding corporations cited AI adoption as a key driver of increased information security investment. The survey covered corporations in 18 countries across industries including IT, manufacturing, finance, and retail and wholesale.
Among the major security threats recognized by corporations, AI vulnerabilities also accounted for a significant share. Nineteen percent of respondents cited AI vulnerabilities as one of the most dangerous threats their organizations face. Phishing was 22%, and large-scale malware attacks were 20%.
In particular, there was strong concern about "shadow AI," in which employees use external AI services such as ChatGPT or DeepSeek during work without company management or control. Thirty-one percent of respondents viewed employees' use of external AI tools as a security risk factor.
However, many corporations chose to set certain limits rather than completely blocking external AI use. Fifty-nine percent of responding corporations allowed the use of external AI tools under conditions such as restricting the upload of company names or internal documents. Twenty-three percent said they allowed use without separate restrictions but provided security guidelines and training.
Despite security concerns, corporations continue to move forward with AI adoption. Among those surveyed corporations, 81% were reviewing or pursuing the introduction of in-house AI tools based on large language models (LLMs), and 18% had already applied related technologies to their business processes.
The main reasons corporations are introducing in-house AI solutions were improved process efficiency, reduced human error, and improved quality of deliverables provided to customers, each at 59%.
Vladislav Tushkanov, group manager at the Kaspersky AI Technology Research Center, said, "As corporations adopt AI to streamline work, security risks such as AI vulnerabilities and shadow AI are growing as well," adding, "It is necessary to secure visibility into AI use and manage it by integrating with existing security frameworks."
Lee Hyo-eun, head of Kaspersky Korea, said, "As Korean corporations expand AI adoption, they must respond to data leakage risks from unmanaged external AI use and to AI-based threats," adding, "An approach that considers security from the AI adoption stage is needed."
