Shadow AI Becomes Major Risk for Enterprises

AI adoption has become nearly universal across Latin American workplaces, but EY's Work Reimagined Survey finds that organizations are struggling to govern its use. While 93% of employees already use AI at work, many rely on personal AI tools outside IT oversight, creating "Shadow AI" risks that threaten cybersecurity, compliance, and data governance.

AI is becoming embedded in daily work across Latin America faster than organizations can establish the governance needed to manage it. According to EY's 2025 Work Reimagined Survey, employees are rapidly incorporating AI into their workflows, but many are doing so independently by introducing personal AI applications into corporate environments, exposing businesses to growing security and compliance risks.

The findings suggest that the challenge facing executives has shifted from encouraging AI adoption to governing it effectively. Although AI has become a common workplace tool, most organizations have yet to transform widespread usage into measurable business outcomes through structured implementation, workforce development and governance.

"The real challenge of AI is not only technological, it is organizational. Shadow AI is the clearest sign that talent is already moving ahead and that there is still enormous potential to capture," says Carolina González, People Consulting Leader, EY Latin America.

González also notes that when it is channeled correctly, AI not only drives productivity, it redefines how we work, supports better decision-making and improves performance. 

“But when adoption moves ahead of strategy, that potential is diluted and organizations lose a significant share of the value AI can generate," says González.

AI Adoption Is Outpacing Enterprise Governance

The survey found that 93% of employees across Latin America already use AI in their daily work, exceeding the global average of 88%. Employees report saving an average of nine hours per week through AI-assisted tasks, compared to the global average of eight hours.

However, the research shows that adoption remains concentrated in relatively simple activities. Information searches account for the most common use case, followed by document summarization and email drafting. More sophisticated applications, including deep research, decision evaluation, and AI-powered mentoring, remain far less common. Only 5% of AI users qualify as advanced users who combine multiple AI tools, assistants, and agents to substantially reshape how they work.

This gap between widespread adoption and advanced implementation helps explain why relatively few organizations are realizing significant returns from their AI investments.

Shadow AI Creates New Security and Compliance Risks

EY identifies Shadow AI as one of the most pressing consequences of this imbalance. Depending on the industry, between 27% and 56% of employees in Latin America are using personal AI tools in the workplace without oversight from IT departments. While this demonstrates employees' willingness to innovate and improve productivity, it also introduces new risks related to data protection, cybersecurity, regulatory compliance, and governance.

Rather than viewing Shadow AI solely as a security concern, the study suggests it reflects unmet demand for enterprise AI capabilities. Employees are adopting external tools because they perceive them as more accessible or better suited to their daily work than the solutions formally provided by their organizations.

The report argues that sustainable AI adoption depends on more than deploying new technology. According to EY, organizations that generate stronger productivity gains combine three factors: employee skills, access to appropriate AI tools, and a workplace culture that encourages responsible adoption. Formal incentives, AI-specific learning programs, role-based tools and leadership confidence all contribute to higher AI maturity.

Training also emerges as a significant driver of adoption. Employees who complete more AI learning hours demonstrate substantially higher levels of AI usage, which in turn correlates with greater productivity gains. At the same time, EY highlights a "learning paradox," finding that employees with more than 80 hours of AI training are 55% more likely to leave their employer than average, underscoring the importance of pairing workforce development with broader talent retention strategies.

The study also identifies notable differences in AI readiness across global markets. Asia-Pacific countries dominate EY's AI maturity rankings, while Latin American markets show mixed performance. Brazil records the region's strongest AI score at 41, above the global average benchmark of 34, while Mexico scores 30, placing it below the global average alongside Colombia.

Despite rapid adoption, only 28% of organizations successfully translate AI implementation into tangible business value, according to the survey. EY attributes this gap to weaknesses in talent strategies, organizational culture, continuous learning and governance rather than limitations in the technology itself.

The findings also reveal broad alignment between employers and employees regarding AI-related risks. Both groups identify security, protection, and oversight as their primary concern, followed by the potential erosion of human expertise through excessive dependence on AI. Job displacement, privacy risks, misinformation, and the need for workforce reskilling also rank among the most significant issues associated with expanding AI adoption.

For business leaders, the survey signals that the competitive advantage created by AI will increasingly depend on organizational readiness rather than access to technology. "The difference will lie with those who integrate talent, culture, technology, and governance to scale AI responsibly. In this new reality, AI is not simply another tool. It is the factor that will define leadership and competitive advantage," González says.