AI moves fast. Stay in the know.

A recent PagerDuty survey found that employees are widely using unauthorized AI tools at work and sharing business information through public platforms such as ChatGPT, Claude, and Gemini. Customer data, emails, meeting notes, financial information, and confidential documents are being entered into AI systems without complete organizational oversight. The findings show that restrictive policies alone may not prevent shadow AI when employees believe approved tools do not meet their needs.

Source: TechRadar

What to know:

• PagerDuty surveyed 1,250 office professionals across the US, UK, Australia, and Southeast Asia.
• Two-thirds said they had used AI tools at work despite believing those tools were not permitted by company policy.
• Eighty-eight percent had shared some form of work-related information with public AI platforms.
• Forty-three percent had entered work emails, while 40% had shared meeting notes or summaries.
• Thirty-four percent had shared customer information, and 31% had entered financial data, confidential documents, or business strategies.
• At companies with fewer than 1,500 employees, 40% had shared customer data with public AI tools, compared with 27% at larger organizations.
• Seventy-seven percent believed workplace AI restrictions were limiting their professional development, suggesting that bans alone may encourage employees to bypass approved processes.

Why it matters:

For mid-sized businesses adopting GenAI, shadow AI can make it difficult to know which tools employees are using, what information is being shared, and whether internal policies are being followed. Personal accounts and browser-based AI tools may also remain invisible to traditional software inventories.

This reinforces the need for continuous AI usage visibility, prompt-level monitoring, and sensitive-data protection. Proactive AI risk monitoring can help organizations identify shadow AI activity, understand where sensitive information may be exposed, and enforce safer AI usage without blocking productive adoption.

A recent IBM survey found that CIOs and CTOs are being held accountable for increasingly autonomous AI systems without having the visibility or governance capabilities needed to supervise them effectively. As organizations deploy more AI agents to increase productivity and accelerate operations, existing controls designed for slower, human-led workflows are struggling to keep pace. The findings show that scaling AI without embedded monitoring and governance can increase data exposure, security incidents, compliance failures, and the need for human intervention.

Source: ITPro

What to know:

• IBM surveyed 2,000 C-level technology executives about AI adoption, governance, architecture, and investment readiness.
• Two-thirds of respondents said they are responsible for AI systems they cannot realistically supervise, while only 11% considered their organizations fully prepared for AI agent deployment at scale.
• Seventy-seven percent said AI adoption is already advancing faster than their organization’s governance capabilities.
• Organizations experienced an average of 54 AI-agent incidents during the previous year that involved unintended or harmful outcomes requiring human correction.
• Of the reported incidents, 37% resulted in data exposure or security breaches, 33% caused cascading system failures, and 17% triggered compliance issues.
• Seventeen percent of incidents were classified as high severity and required more than four hours to contain.
• Organizations that embedded controls directly into their AI systems experienced 25% fewer incidents than those relying mainly on manual governance processes.
• Nearly six in ten technology leaders identified security and compliance concerns as major barriers to scaling AI agents.

Why it matters:

For mid-sized businesses adopting GenAI, the pressure to improve productivity and deploy AI faster can result in systems being introduced before sufficient visibility, monitoring, and governance are in place. When AI agents operate continuously across applications and business data, manual reviews may not detect risky prompts, sensitive data exposure, policy violations, or unintended actions quickly enough.

This reinforces the need to build AI usage visibility and governance into adoption from the beginning. Continuous monitoring, prompt-level insights, sensitive-data protection, policy-aligned guardrails, and auditability can help businesses understand how AI is being used, identify risks as they emerge, and scale GenAI without allowing productivity gains to create unmanaged security, compliance, or operational exposure.