The use of AI within organizations is growing rapidly, but control over it is lagging behind. A recent global survey by KPMG on trust and attitudes toward AI use found that 58% of employees regularly and consciously use AI tools in their work. Of these, 70% use freely available, public AI tools. Nearly half of these users admitted to entering company information—such as financial data, sales figures, or customer details—potentially exposing their employer to risks like data breaches, reputational damage, and non-compliance.
Technology is not standing still—far from it. Experts expect a widespread breakthrough of so-called AI agents in 2025: autonomous systems that perform tasks, make decisions, and communicate with other software. These tools will unlock many new use cases and productivity gains, but they also increase the overall risk profile of organizations—with complex risks such as excessive autonomy and the loss of human oversight. Without targeted investment in governance, control mechanisms, and the development of AI skills, it will become increasingly difficult to use this technology responsibly.
Yet the same survey shows that nearly two-thirds of employees have never received any AI training. This creates a worrying gap: AI use is growing, but ‘AI literacy’ is lagging—despite new European AI legislation requiring organizations to provide appropriate training and awareness around AI use, especially for high-risk applications. In terms of governance and oversight, many organizations are still underprepared. Only 55% of employees in advanced economies believe their organization has sufficient guidelines, controls, or safeguards in place to ensure responsible AI use. The future of work is changing rapidly, and organizations must adapt to remain both relevant and secure.