BlackFog launches ADX Vision to combat rising shadow AI risks

BlackFog has launched ADX Vision, a new software tool designed to give enterprises greater oversight and control over employees' use of unapproved artificial intelligence technologies, a practice known as shadow AI.

Shadow AI risk

Shadow AI refers to the use of AI tools by employees without the approval or oversight of IT or security departments. This can create openings for the unintentional loss or leakage of sensitive company data and intellectual property.

BlackFog's move follows company-sponsored research that surveyed 2,000 employees across the United Kingdom and United States. The findings reveal that 49% of employees use AI applications at work without their employer's approval. There are additional gaps in understanding; just 53% of respondents say they know how input data is saved, analysed, or retained by these third-party AI tools.

Despite these risks, the survey also found that 71% of employees believe that the productivity gains from unauthorised AI tools outweigh the data privacy concerns.

Data governance

ADX Vision operates by monitoring endpoint devices for unauthorised use of AI tools, and is designed to detect suspicious activity in real-time. By enforcing data governance policies directly on the user's device, BlackFog says the solution can help organisations prevent the exfiltration of sensitive or regulated information while minimising disruption to everyday workflows.

The platform notifies administrators when unsanctioned AI applications are used and blocks the movement of certain types of information, supporting compliance with regulatory frameworks around personal and business data.

Real-time controls

ADX Vision is initially being released for Microsoft Windows, with future support for macOS and Linux expected in early 2026. The software builds on BlackFog's existing anti-data exfiltration suite, with a specific focus on the risks introduced by the adoption of large language models and other AI-based productivity tools in the workplace.

Unauthorized tools can learn from, store, and transmit sensitive enterprise information outside company firewalls, raising exposure to financial and reputational harm. BlackFog's approach centres on providing enterprise IT teams with dashboards and alerts to review and stop suspicious usage before data leaves a device.

Industry concerns

"As AI adoption accelerates, so does the risk of data loss through unmonitored tools. This research highlights the extent of that risk and the critical need for organizations to protect the data flowing through AI systems," said Dr Darren Williams, CEO and Founder, BlackFog.