Mexico City, Mar 10 (EFE).- The accelerated adoption of artificial intelligence (AI) is creating critical challenges for global cybersecurity, according to the report “Fault Lines in the AI Ecosystem” prepared by TrendAI, which recorded an increase of nearly 35% in 2025 compared with the previous year.
The report revealed that in 2025 there were 2,130 vulnerabilities linked to AI, a 34.6% increase from the previous year, representing 4.42% of all software flaws detected worldwide.
The analysis warned that, if the trend continues, the number of vulnerabilities related to artificial intelligence could range between 2,800 and 3,600 cases in 2026, consolidating it as one of the main risks in the global cybersecurity landscape.
TrendAI Senior Vice President Hernan Armbruster stressed the importance of adopting comprehensive strategies to ensure the security of AI systems.
“Artificial intelligence security must be a strategic priority for organizations,” he said.
The report identified several “fault lines” where vulnerabilities are increasing, including exposed infrastructure, insecure data channels, model weaknesses, and governance gaps that adversaries are increasingly exploiting.
Specialists warned that many organizations are rapidly adopting AI tools to improve productivity, automate processes, and strengthen competitiveness, but without full oversight of the systems they use.
One concerning phenomenon identified by the study is the use of unauthorized implementations known as “shadow AI,” which lack adequate oversight and security controls, increasing the risks of data theft and malicious manipulation of AI models.
AI also empowers attackers
In addition, the report noted that attackers are using AI tools to automate vulnerability reconnaissance, refine phishing campaigns, and scale digital fraud operations, increasing the speed and sophistication of attacks.
“In today’s environment, attackers use AI to improve precision and scalability,” explained Resecurity CEO Gene Yoo, who emphasized that organizations will need to respond with equally advanced defenses.
Among the most alarming findings of the report is the growth in vulnerabilities within the large language model (LLM) ecosystem, which increased by 80.4% over the past year, reflecting the rapid expansion of these tools in corporate environments.
According to the analysis, more than 50% of the vulnerabilities detected in the AI ecosystem are related to specialized hardware, such as GPUs and accelerators used to train and run advanced models.
This type of infrastructure has become an attractive target for attackers seeking to compromise entire enterprise systems or gain access to large volumes of data.
SISAP CEO Mauricio Nanne emphasized that harnessing the potential of artificial intelligence will require more secure technological architectures, specialized cybersecurity talent, and proactive threat detection strategies.
“AI represents an extraordinary promise, but also a significant responsibility,” he said.
Global debate on AI security
The growth of risks associated with artificial intelligence is fueling an international debate on the governance of this technology and the protection of digital infrastructure. Experts anticipate that issues such as international cooperation and digital resilience will take center stage at upcoming meetings in the technology sector.
Platforms such as Digi Americas Alliance and the Digi Americas LATAM CISO Summit are emerging as key spaces to address these challenges. At these gatherings, public and private sector leaders will analyze how to strengthen regional collaboration in the face of increasingly sophisticated digital threats. EFE