TrendAI has published research examining global AI deployment alongside associated security and compliance risks. The study surveyed 3,700 business and IT decision makers.
The findings show that 67% of respondents feel pressure to approve AI projects despite security concerns, driven by competitive factors and internal demand. Some respondents described these concerns as “extreme,” but still proceeded with deployment.
The research also highlights governance challenges, including inconsistent policies and unclear accountability for AI-related risk. In some cases, security teams are required to respond to AI deployment decisions after they have been made, which can result in workarounds and the use of unauthorised or “shadow” AI tools.
Additional TrendAI threat research indicates that attackers are using AI to automate activities such as reconnaissance and phishing, increasing the speed and scale of cyber threats.
The study suggests that AI adoption is progressing faster than organisations’ ability to manage associated risks. Around 57% of respondents said AI is advancing more quickly than they can secure it, while more than half reported only moderate confidence in their understanding of relevant legal frameworks.
Governance maturity remains limited, with approximately 38% of organisations reporting comprehensive AI policies in place, while others are still developing them. Around 41% cited unclear regulation or compliance standards as a barrier.
Confidence in autonomous AI systems is still developing. Fewer than 44% of respondents believe agentic AI will significantly improve cyber defence in the near term. Reported concerns include data access (42%), malicious prompts (36%), and risks related to expanded attack surfaces and autonomous code deployment (33%).
The research also highlights challenges in visibility and auditability, with 31% of organisations reporting limited ability to monitor these systems. Around 40% support the introduction of AI “kill switch” mechanisms, while others remain uncertain.
Overall, the findings indicate that organisations are continuing to expand AI adoption while governance, security, and oversight practices are still evolving.
