AI is speeding up the development of Web services, but security flaws are emerging, exposing weaknesses in how Japanese companies use AI. Addressing the vulnerabilities that have been identified is likely to be an issue going forward.
AI can generate code instantly and dramatically increase development speed, but this has raised the question of who will take responsibility for the security of that code. Companies are entering an era in which they are being judged not only on their ability to use AI, but also on their ability to use data securely, and the development of data infrastructure designed for AI is increasingly being prioritized.
However, many companies have fallen behind in establishing data management infrastructure and access controls. There are numerous cases in which companies lack an adequate understanding of where data is located, who manages it and who has access, raising concerns that the scope of AI’s permissions may diverge from the intended purpose of its actual use.
With the emergence of frontier AI, the number of vulnerabilities being discovered is trending upward, and cases have been reported in which autonomous systems identify vulnerabilities and go on to develop and execute attacks and steal data. The automation of attacks using generative AI is advancing, creating situations that are difficult to address through conventional operations alone.
In response, companies around the world are accelerating the shift away from excessive reliance on SIEM-based monitoring toward AI-enabled automated analysis and data-centric defense. Competitiveness in the AI era will depend not only on adopting superior AI models, but also on establishing the data infrastructure and governance needed to create an environment in which AI can operate safely.