Uncontrolled use of AI in organisations linked to rising risk and slower returns

New polling from Lenovo suggests that the widespread, and often unregulated, use of artificial intelligence in organisations is creating operational risks, increasing costs and slowing the return on investment from AI initiatives. The company’s latest Work Reborn Report, based on a survey of 6,000 employees worldwide, claims that more than 70 percent of employees now use AI tools on a weekly basis. Up to a third of this activity is taking place without formal oversight from IT departments, contributing to the growth of so-called shadow AI.

The findings point to a widening gap between the adoption of AI and organisations’ ability to manage it effectively. While 80 percent of respondents expect to increase their reliance on AI over the next year, governance and control mechanisms are not keeping pace. This has implications beyond IT functions, particularly for security leaders who must manage expanding risks across devices, endpoints and data flows.

According to the report, 61 percent of IT leaders have already seen an increase in cybersecurity threats linked to AI. However, only 31 percent say they feel confident in their ability to manage these risks. Concerns are also reflected among employees, with 43 percent expressing worries about data exposure or AI-related attacks.

Lenovo argues that the lack of oversight is already affecting business performance. Fragmented adoption of AI across teams is associated with delayed returns, duplicated spending on overlapping tools and limited visibility into which systems are delivering value. The report also highlights uneven access to AI across the workforce, with some employees operating within managed environments while others rely on unapproved tools to maintain productivity. This disparity is described as contributing to a “two speed” workforce, potentially slowing decision-making and increasing inefficiencies.

The research suggests that many organisations are attempting to manage AI across disconnected systems, with devices, infrastructure and security handled separately. This fragmentation is identified as a key factor behind the gap between AI usage and effective execution.

Lenovo’s report outlines an alternative approach centred on integrating device management, infrastructure and security into a single operating model. It argues that embedding control at the device level, combined with ongoing managed security services, could help organisations reduce complexity and improve oversight.

The findings reflect broader concerns about how organisations govern emerging technologies, particularly as adoption accelerates faster than internal systems and policies can adapt.