Your IT Team Is More Open to AI Security Tools Than You Think — With One Catch

That gap — between stated trust and actual behavior — is the most revealing finding in OpenVPN's new research on how IT and business leaders think about AI in their security stack. And it has direct implications for how organizations should evaluate, position, and adopt AI-powered network security tools.

The trust number is lower than you'd expect

OpenVPN recently surveyed 45 IT and business leaders about their attitudes toward AI in security tools. When asked how comfortable they are with AI assisting in cybersecurity tasks, the results were more measured than the industry conversation might suggest.

Only 5% said they fully trust AI to assist in security decisions.

The dominant response — by a wide margin — was "somewhat comfortable, with human oversight required." That framing matters. It's not resistance to AI. It's a condition placed on it.

A further 25% said they were neutral or unsure, and a combined 16% expressed some level of discomfort. The picture that emerges is an audience that hasn't written off AI, but hasn't fully bought in either — and is watching carefully for the right use case.

So why are 68% ready to use AI features right now?

Here's where it gets interesting.

Despite only 5% expressing full trust in AI for security decisions, 68% of respondents said they'd try or actively use AI features if their security platform offered them.

That's not a typo. The same audience that's skeptical of AI autonomy is, in the majority, willing to use AI capabilities in practice.

The explanation lies in what those capabilities actually look like. When AI is framed as a feature — something that surfaces anomalies, flags unusual access patterns, suggests policy adjustments — it's a tool in the hands of a human decision-maker. That's fundamentally different from AI acting as the decision-maker itself.

The research didn't ask respondents to trust AI. It asked whether they'd use it. Those are different questions, and the answers are different for a reason.

The neutral segment is the biggest opportunity

General AI sentiment in the survey broke down as mostly positive — 56% hold a positive view of AI in business software overall. But the 23% who said they were neutral deserve attention.

This isn't the same as skepticism. Neutral respondents haven't seen a compelling enough use case to form a strong opinion either way. They're not resistant — they're unconvinced.

For security vendors and IT leaders evaluating tools alike, this is the most actionable segment. A well-positioned AI feature — one that demonstrably reduces noise, surfaces real signals, or saves meaningful time — can move a neutral to a positive without requiring a trust argument at all.

The data suggests the path to AI adoption in security isn't convincing people to trust AI more in the abstract. It's showing them a specific capability that makes their work better.

What this means for evaluating AI-powered security tools

If you're an IT leader currently evaluating network security platforms that include AI features — or deciding whether to enable them — the research offers a useful framework.

• Look for assistive, not autonomous. The strongest signal in the data is that IT teams want AI in the loop, not in the seat. Tools that position AI as "it tells you, you decide" will land better than tools positioning AI as "it decides, you're notified."
• Comfort levels differ by task type. The survey found that general AI sentiment is more positive than cybersecurity-specific AI sentiment — suggesting that the bar is higher when decisions carry security risk. Evaluate AI features by their specific function, not by general AI reputation.
• Even skeptics convert. Among respondents who expressed negative views of AI in business software generally, a meaningful portion still said they'd be likely to use AI features in a security platform. Context changes the equation.

The full picture is in the report

The AI findings are one section of a broader research report OpenVPN published on access security, Zero Trust adoption, and AI readiness among IT and business leaders. The complete data includes the full breakdown of AI sentiment by segment, the adoption and hesitation data around Zero Trust, and six specific recommendations for IT teams evaluating their network access strategy.

If your organization is currently assessing AI-powered security tools — or building the internal case for Zero Trust — the report is worth the download.