AI Security Concerns: Internal Threats, Obsolete Protection Measures, and Complexity of Artificial Intelligence
In a world where artificial intelligence (AI) is increasingly being adopted, a new study commissioned by Opswat reveals that businesses are facing significant AI-based file security risks. The study highlights that only 25% of companies have a GenAI policy, and a staggering 61% have experienced file breaches due to negligent or malicious insiders in the last two years.
The study underscores the importance of unified multi-layered platforms for addressing these risks. George Prichici, Vice President of Products at Opswat, emphasizes the significance of these platforms, stating that they enable file security architectures to be flexibly adapted to new threats and effectively protect modern workflows and complex file ecosystems within and beyond the perimeter.
The report reveals that multi-layered defense mechanisms, such as zero-trust file handling combined with prevention tools, are now considered an indispensable standard for robust and scalable security in the AI era. However, these technologies, including multiscanning, content disarm and reconstruction (CDR), and sandboxing, are far from ubiquitous, with only 73% of companies planning to adopt them by 2026.
The study also sheds light on other areas of concern. For instance, only 40% of companies manage to detect file-based threats within a week, and file sharing and transfers are identified as weak points, with only 39% of respondents expressing confidence that file transfers to third parties are reliably protected.
Incomplete security of AI workloads is another area of concern, with 37% securing sensitive data in AI workflows using fast security tools, masking, or within guardrails. Macro-based and Zero-Day malware are identified as the most concerning threats, with 44% naming macro-based malware as the greatest threat.
The full study, including more detailed findings, is available for download. It is clear that in the AI era, businesses must prioritise the adoption of unified multi-layered platforms to ensure robust and scalable security. As AI adoption continues to grow, so too must governance and security measures to mitigate the risks associated with AI-based file security.
