According to the recently published Thales 2026 Data Threat Report, many organizations say that the rapid pace of AI-driven transformation is now their biggest security challenge. Based on the report’s research, conducted by S&P Global 451 Research, 61% cite AI as their top data security risk. The concern is not only about malicious AI, but about the access it is being granted as it shifts from a tool to a trusted insider.
As enterprises embed AI into workflows, analytics, customer service, and development pipelines, these systems are being granted broad, automated access to enterprise data, often with fewer controls than those applied to human users in a corporate environment.
“Insider risk is no longer just about people. It is also about automated systems that have been trusted too quickly,” says Sebastien Cano, Senior Vice President, Cybersecurity Products at Thales. “When identity governance, access policies, or encryption are weak, AI can amplify those weaknesses across corporate environments far faster than any human ever could.”
The report reveals a disconnect between AI adoption and data control. Only 34% of organizations know where all their data resides, whatever the level of criticality, and just 39% can fully classify it. Meanwhile, nearly half (47%) of sensitive cloud data remains unencrypted.
As AI systems ingest and act on data across cloud and SaaS environments, limited visibility makes enforcing least-privilege access increasingly difficult, that is, granting only the strictly necessary access rights. This increases the extent of exposure if credentials are compromised.
Identity infrastructure is now the primary attack surface. Credential theft remains the leading attack technique against cloud management infrastructure, cited by 67% of organizations experiencing cloud attacks. At the same time, 50% rank secrets management among their top application security challenges, reflecting the growing complexity of governing machine identities, API keys, and tokens at scale.
AI is powering more convincing attacks
While organizations race to adopt AI, attackers are doing the same. Nearly 60% of companies report experiencing deepfake-driven attacks, and 48% report reputational damage tied to AI-generated misinformation or impersonation campaigns.
As AI introduces new risks, it also increases existing ones. Human error already contributes to 28% of breaches, and with automation layered on top, small mistakes can scale faster and spread wider.
Security investment is developing, but not at the pace of the new risks
While organizations recognize the need to adapt, investment is not keeping pace with the rapid expansion of AI-driven access and automation. 30% now dedicate specific budgets to AI security, reflecting growing awareness. However, the majority (53%) still depend on traditional security programs built primarily for human users and perimeter-based controls. As machines increasingly authenticate, access, and act autonomously, many security strategies have yet to adjust to this shift in operating models.
