IBM and security experts address AI integration and security risks

As artificial intelligence becomes integrated into business operations, experts are highlighting the need for robust security strategies to manage emerging threats. A significant concern is ‘Shadow AI’—the use of unauthorized AI services by employees. According to Gartner, while 75% of Japanese companies allow user-selected generative AI, 73% struggle to effectively manage Shadow AI, risking data leaks, compliance violations, and intellectual property issues.

To address the growing ‘AI divide,’ IBM has launched ‘IBM Enterprise Advantage’ in Japan. This platform aims to help organizations redesign business processes to work alongside AI agents. IBM reports that within its own operations, it has deployed over 4,000 digital workers to assist human teams.

In the realm of cybersecurity, experts advise against fully delegating complex tasks to AI. Instead, they recommend starting with repetitive, high-volume tasks—such as alert triaging or vendor risk assessments—to reduce fatigue while maintaining human oversight. This approach helps mitigate risks like ‘ghostjacking,’ where attackers embed malicious instructions into data read by AI agents, and ensures that AI-generated patches do not introduce new vulnerabilities.