AI has moved from experimental to essential in many enterprise environments. Still, a recent workplace sentiment survey conducted by WitnessAI reveals an emerging challenge: employees are adopting AI faster than organizations are securing it.
The survey found both encouraging and cautionary: AI is primarily viewed as productive but unprotected.
Productivity Wins—But With Risk
62% of respondents described their experience with AI tools as empowering, saying they make them more productive. Another 59% said AI saves time and reduces busy work. These responses show a workforce that isn’t skeptical about AI’s role; it’s embracing it.
However, when asked what concerns them most, 43% cited “accidentally sharing sensitive data” as their top worry. That was more than double the next-highest response (job displacement, at 14%).
This isn’t fearmongering; it’s rational concern. AI tools often operate with vast access to unstructured data. Without clear parameters, it’s easy to see how sensitive information can slip through the cracks.
A Policy Vacuum
Even more striking: only 17% of respondents said their organization has clear policies and training around AI use. The majority (43%) said the rules are clear, while 24% said there are no guardrails.
That’s a red flag for CISOs, CIOs, and compliance leaders. AI tools are being used—and employees are engaging with them—but they’re doing so in an environment lacking structure. This opens the door for accidental data leakage, regulatory violations, and reputational harm.
What the Workforce Wants
Critically, this isn’t a call to slow down. It’s a request to move forward responsibly. When asked what would make them feel safer using AI at work:
- 35% said human oversight of AI decisions
- 31% want access to training and support resources
- 17% called for dedicated AI governance teams
These findings reveal a workforce that’s not just open to AI—it’s ready for it. But they want protections to ensure it’s used safely, transparently, and compliantly.
From Adoption to Enablement
This moment represents a pivotal shift. The conversation is no longer about whether employees will accept AI. They already have. The next phase is whether organizations can implement AI securely, with built-in policy enforcement, real-time oversight, and visibility into how models interact with sensitive data.
The opportunity is clear: build AI systems that are secure by design, align with employee expectations, and restore confidence in the speed of innovation.
Join our LinkedIn group Information Security Community!
