By Stephanie Simone
Public company disclosures of AI-related risks have surged from12% of S&P 500 firms in 2023 to 72% in 2025,underscoring the urgency for stronger security and governance strategies as AI adoption continues to accelerate.
Organizations must now secure not just data and infrastructure, but also models, pipelines, and autonomous AI behavior. The challenge is balancing rapid innovation with security, compliance, and control.
DBTA held a roundtable webinar, Taming AI Risks: Strategies for Security and Governance, featuring experts in the field who explored how leading organizations are identifying, managing, and mitigating AI-related risks.
Dany Nehme, director, solutions engineering at insightsoftware, cited several statistics including “74% of companies haven’t seen value from AI.”
Connecting to live data sources is difficult and dangerous. AI agents need access to live data for context and accuracy, Nehme said.
AI needs a trusted layer to be production ready. With Simba Intelligence, trusted data and AI finally connect, he noted.
According to David Thain, product marketing lead at Informatica, enterprise-wide AI governance is difficult because of siloed expertise, fragmented data, slow processes, and evolving standards.
However, governance makes AI worth deploying. Use governance principles to start taming AI risks. According to Thain, this includes:
- Lineage: Lineage is the backbone of explainability
- Classification: Bias enters through data, not just algorithms
- Access Control: Govern what AI can see and when
- Data Quality: Quality issues become AI reliability failures
For trusted agents, make sure you have trusted data. Identify and resolve issues in data pipelines. Understand data and AI assets to create trust. Control access to data before sharing. And package models for easy consumption.
James Willis, director, solutions consulting at OpenText, said that to be ready for AI, enterprises should:
- Discover: Know what content you have across every system and format
- Prepare: Make it understandable and usable for GenAI
- Govern: Keep it governed, compliant, and protected at scale
- Action: Activate AI outputs into business impact
For the full webinar, featuring a more in-depth discussion, Q&A, and more, you can view an archived version of the webinar here.