A new global study by TrendAI shows that many organizations are rolling out artificial intelligence (AI) even when they have serious security and compliance concerns.
The survey covered 3,700 IT and business decision-makers worldwide, including 200 in the Philippines. It found that 67 percent of IT leaders and 72 percent of business leaders in the country felt pressured to approve AI projects despite risks. About one in seven said their concerns were “extreme” but were set aside to keep up with competitors and internal demand.
Most organizations said they are trying to protect their systems, with 76 percent focusing on data security and 72 percent on data quality. However, experts say pressure to adopt AI is outpacing proper safeguards.
Rachel Jin of TrendAI said companies understand the risks but lack the right systems to manage them. She warned that rushing AI into critical operations without strong controls could create bigger problems.
The study also found weak governance and unclear responsibility for managing AI risks. Security teams often react after decisions are made, leading to the use of unauthorized or “shadow” AI tools.
At the same time, cybercriminals are increasingly using AI to speed up attacks such as phishing and reconnaissance, making threats more frequent and harder to stop.
Only 33 percent of IT leaders and 44 percent of business leaders said they feel well prepared for the rapid growth of AI. While more than half of organizations already have AI policies, about one-third are still developing them. Key challenges include unclear regulations and a lack of skilled experts.
Confidence in more advanced AI systems remains mixed. While 60 percent believe autonomous or “agentic” AI could improve cybersecurity, many worry about risks such as access to sensitive data, misuse, and lack of oversight.
More than half of organizations said their top concern is AI systems accessing sensitive information. Others pointed to risks like malicious inputs, abuse of AI privileges, and a growing number of cyberattack opportunities. Around 30 percent admitted they cannot fully monitor or audit these systems.
In the Philippines, 65 percent of organizations support having an emergency “kill switch” to shut down AI systems if something goes wrong, while 33 percent are unsure.
The findings highlight a growing gap between fast AI adoption and the ability to manage its risks, raising concerns about how organizations can stay in control as the technology advances.
