Skip to content

Content Library

Back to Articles

Article August 26, 2026

Why AI Makes Third Party Risk a Boardroom Imperative

Why AI Makes Third Party Risk a Boardroom Imperative

Artificial intelligence is transforming third-party risk from a compliance function into a strategic capability. The SVP, Head of Audit for a major international bank explains why continuous monitoring, ecosystem thinking and stronger governance are becoming essential as AI reshapes vendor risk across the financial services sector.

  • AI has made third-party risk unavoidable because banks increasingly depend on AI-enabled vendors
  • Traditional periodic vendor reviews are giving way to continuous intelligence-driven monitoring
  • Third-party risk is expanding into fourth-party and wider ecosystem risk with shared technology dependencies
  • Explainability, model drift, hallucinations and data governance create major new AI oversight challenges
  • The three lines of defense remain relevant but require broader specialist expertise across multiple risk domains
  • Stronger contracts, continuous monitoring and coordinated governance will define future third-party risk management

Banks can no longer treat third-party risk management as a compliance exercise focused on periodic vendor reviews.

As artificial intelligence becomes embedded throughout the technology supply chain, organizations must fundamentally rethink how they identify, monitor and govern external risk.

Speaking at CeFPro's 2026 Risk Americas conference the SVP, Head of Audit for a major international bank argued that institutions should transform third-party risk management into a strategic capability.

This, he said, should be supported by continuous monitoring, specialist expertise and AI-powered intelligence if they are to remain resilient in an increasingly interconnected ecosystem.

The executive explained that AI presents a unique challenge because organizations cannot simply choose to avoid it.

"We cannot say we're not going to use this vendor because it is AI enabled. You don't have that option available," he said. "So you have no choice but to use vendors with AI embedded in them."

As a result, AI itself becomes an unavoidable source of third-party risk that institutions must understand and manage rather than avoid.

He argued that this represents a decisive shift in the purpose of third-party risk management.

Historically, vendor oversight relied heavily on spreadsheets, manual processes and reactive intervention after problems had already emerged.

Following the financial crisis, firms developed more formal governance, centralized vendor inventories and periodic monitoring aligned with regulatory expectations.

Today, however, those approaches are proving inadequate for an environment where AI-enabled vendors, cloud providers and complex technology ecosystems evolve continuously.

The speaker described the industry as approaching an inflection point where continuous intelligence replaces periodic assessment.

Instead of annual reviews, institutions should increasingly rely on integrated internal and external data feeds, financial risk indicators, cyber intelligence, adverse media and operational signals to produce near real-time vendor risk scoring and automated alerts.

Looking further ahead, he envisaged AI eventually predicting vendor failures, cyber incidents and compliance breaches before they materialize, ultimately creating self-optimizing third-party risk frameworks capable of monitoring vendor onboarding, performance and remediation across entire supply chains.

That evolution, however, introduces an entirely new collection of interconnected risks.

Traditional operational risk is expanding into model and automation risk. Cyber threats increasingly involve adversarial AI.

Data protection has become inseparable from AI training practices, while third-party risk now extends well beyond immediate suppliers into fourth-party and wider ecosystem dependencies.

The executive warned that organizations are becoming increasingly dependent on shared cloud infrastructure, common AI models and a relatively small number of technology providers.

"One vendor relying on the same AI processor or multiple vendors subject to the same geopolitical risk" can create cascading failures that rapidly spread across interconnected organizations, he explained.

Understanding those relationships, rather than assessing vendors individually, will become one of the defining capabilities of effective third-party risk management.

Model risk also takes on greater significance when AI becomes embedded within vendor products.

Black-box algorithms, limited explainability, model drift and hallucinations all create the potential for inaccurate decisions, regulatory breaches and diminished auditability.

The executive explained that explainability ultimately comes down to answering straightforward but essential questions:

"Why did AI make this decision? What factor influenced the outcome? How confident is the model? Can humans understand and challenge the results?"

Those expectations mirror long-established model risk management principles while introducing additional complexity unique to AI.

Data governance presents equally significant challenges. Firms must understand precisely where vendor data originates, how it is used to train AI models and whether sensitive information could leak during training or operation.

Cross-border data transfer restrictions, weak data lineage and poor-quality information all increase operational, regulatory and reputational risk.

"Any model, whether it's an AI model or any model, it's as good as the data," he said. "Garbage in, garbage out."

Events you may be interested in

View all of our upcoming events

Sign Up to Our Newsletter

Sign up

Connect Plus

Unlock more with Connect Plus membership

This content is available to Connect Plus members.