Security, Spend and Skepticism: What's Really Holding Back AI in Risk Management

display-middle-ads-here

Despite growing interest in AI-driven solutions, risk professionals are encountering a stubborn set of roadblocks that need to be navigated if technology is to be integrated effectively into the risk management landscape.

In its 2025 white paper, Bridging the Productivity Gap in Third-Party Risk Management with AI, commissioned from CeFPro, Certa identified three major hurdles to adoption: data security, costs, and a lack of internal readiness.

Sixty percent of survey respondents claimed data security is their top concern – hardly surprising, perhaps, in light of high-profile breaches involving AI platforms that have left firms understandably cautious.

The idea that AI systems could be manipulated, or expose sensitive data, is naturally a genuine worry, with experts pointing to the unpredictability of self-modifying systems and a lack of transparency around how data is handled.

Concerns around third-party data access, model drift, and regulatory noncompliance are not just theoretical – they are now routinely included as top agenda items by boards and compliance committees alike.

For firms handling sensitive financial or personally identifiable information, one breach could equate to years of lost trust.

Cost is the next big barrier, with 47 percent of participants citing the high price of implementation as a key reason they’re delaying AI deployment.

And the problem isn’t just upfront investment. More than half of those surveyed reported lack of time and resources, and 50% admitted they were struggling with inadequate tools and existing systems that were running under stress.

Some institutions also face hidden costs linked to integration, training, and ongoing monitoring – costs that often go unaccounted for during the budgeting process, and may not be visible on the bottom line.

Smaller firms, in particular, struggle to justify large-scale AI deployments when operational efficiency is already stretched.

Then comes the skills gap. A third of firms don’t have the personnel needed to operate AI tools effectively, which creates a double-bind: firms know they need AI to improve efficiency, but can’t justify adoption without confidence in the successful rollout.