Research from TrendAI™ shows that the pressure to implement AI for more efficient work is stronger than that for governance, transparency and accountability
STOCKHOLM, March 25, 2026 – The global leading security company TrendAI™ has published a new survey showing that companies worldwide continue to push the adoption of AI, despite known risks associated with both security and regulatory compliance.
The investigation, which has surveyed 3,700 business and IT decision-makers, shows that 67 percent have felt pressured to adopt AI, despite concerns about security. One in seven describe their concerns as ”extreme” but have overridden the feeling because they felt pressured to keep up with competitors.
Companies are aware of risks, but lack the ability to manage them, says Martin Fribrock, Country Manager Sweden, Finland and Baltics at TrendAI. When implementation is driven by competitive pressure rather than maturity in governance and control, we create situations where AI is built into critical systems without the tools needed to manage it safely. This research underscores the importance of helping companies drive strong business outcomes with AI, without compromising risk management.
The risks of letting competition drive AI adoption are compounded by the lack of governance and unclear accountability for AI risks, which are also becoming increasingly common. The same applies to security teams, who, because AI implementation is top-down, are forced to work reactively to management decisions. This often leads to emergency solutions and increased use of shadow AI and other AI tools that are not approved by the organization.
New research from TrendAI™ At the same time, it shows a clear shift: attackers are already using AI to automate their mapping, accelerate phishing, and generally lower the threshold for cybercrime. This is both increasing the speed and scope of attacks.
The control system cannot keep up with AI usage
Companies are implementing AI faster than they can manage the risks, widening the gap between ambition and control. 57 percent say AI implementation is moving faster than they can secure its use, while more than half (64 percent) state that they have only a moderate understanding of the legal frameworks governing AI.
The level of maturity for governance remains low. Only just over a third (38 percent) of companies report that they have comprehensive AI regulations in place, while many are still in the process of developing them. 41 percent believes that unclear regulations or compliance requirements are an obstacle. In practice, this means that AI is introduced into businesses without fully establishing rules for how it should be used.
Trust in autonomous AI remains low
Trust in more advanced, autonomous systems is still in a maturing phase, with fewer than half (48 percent) believing that AI agents will significantly improve cyber defenses in the short term, while concerns about data access, misuse, and lack of control remain.
The data also shows where the concerns are greatest. More than four in ten companies (44 percent) say that what they perceive as the biggest risk is AI agents gaining access to sensitive data. Over a third (36 percent) warn that malicious prompts can compromise security, while one in three (33 percent) point to the growing attack surface for cybercriminals. As many (33 percent) fears abuse of trusted AI tools, as well as risks associated with autonomous distribution of code.
At the same time, almost a third think (31 percent) that they lack transparency or traceability in the systems, which raises serious questions about how organizations can control or intervene once agents are deployed.
Around 40 percent of companies support the introduction of AI-based kill switches to shut down systems in the event of failure or misuse, while almost half are unsafe. The lack of consensus highlights a deeper problem: Companies are moving towards autonomous AI without agreement on how to maintain control when it really matters.
Agentic AI poses a new type of risk for businesses, says Martin Fribrock. Our research already shows clear concerns, from exposure of sensitive data, to loss of control. Without transparency and governance, companies are implementing systems they neither understand nor can control. This risk will only increase if we do not take action.
About TrendAI™
TrendAI™, a business unit within Trend Micro and a global leader in AI security, making the world safer for digital information exchange between enterprises, governments, and organizations. With security expertise and innovation, TrendAI™ leverages artificial intelligence to protect over 500,000 businesses and millions of individuals across AI, cloud, network, and client devices. The flagship TrendAI Vision One™ is the only AI-powered enterprise cybersecurity platform that centralizes cyber risk management and security posture, delivering robust protection across all layers of on-premises, hybrid, and multi-cloud environments. The platform is powered by threat intelligence and insights collected from around the world, helping to protect organizations from hundreds of millions of threats every day. With 6,000 security experts in 75 countries, TrendAI™ empowers enterprises to stay ahead of threats and achieve proactive security outcomes across the entire attack surface, including in mission-critical environments like AWS, Google, Microsoft, and NVIDIA, AI without fear.
Read the full report here.
Swedish
Danish
Norwegian
Finnish
English
German
Dutch