The world of artificial intelligence (AI) is rapidly evolving, and with it, came a new set of challenges. How do you secure workflows that involve humans using AI to access sensitive data? This seems like a simple question, but the answer is far from straightforward. David Goldschlag, Co-Founder and CEO of Aembit, recently broke down the complexities of implementing Identity and Access Management (IAM) for AI at Identiverse 2026.
AIs are not like traditional machines or human beings; they don’t quite fit into our existing security frameworks. They require a unique approach to IAM that takes into account their abilities, limitations, and behavior. Goldschlag’s presentation highlighted the challenges of securing workflows that involve AI agents accessing sensitive data on behalf of humans.
Implementing IAM for AI is not just about granting access to machine-readable data; it involves understanding how to attribute actions, ensure accountability, and maintain transparency in a way that makes sense to both humans and machines. The goal is to enable AI agents to work at scale while maintaining the security standards we expect from human interactions.
Goldschlag explained that IAM systems need to be designed with the specific requirements of AI systems in mind. This includes supporting advanced access control models that can handle the complex relationships between humans and AI entities. He also emphasized the importance of understanding how AI systems use identity, including what constitutes an “identity” in the context of machine-based workflows.
An example discussed by Goldschlag was a real-world implementation where IAM was used to secure data access for human-AI collaborative workflows. The solution involved not only securing the interactions between humans and AI but also ensuring that AI agents could take on roles similar to those of humans, without sacrificing security or compromising data confidentiality.
While it may seem daunting, Goldschlag’s presentation offered practical advice on how businesses can approach IAM for AI. He suggested starting with a thorough assessment of the specific needs of each use case and the types of identities that need to be supported. This includes understanding what attributes are required for authentication and authorization, how identity changes over time, and what constitutes attribute-based access control in the context of AI workflows.
The Identiverse 2026 presentation serves as a reality check for businesses looking to deploy IAM for AI at scale. With the increasing dependence on technology to perform complex tasks, understanding how to secure these interactions has become a critical component of modern business operations. By adopting an integrated approach that bridges human IAM capabilities with those designed specifically for machine-based systems, organizations can better mitigate risks and strengthen their overall security posture.
Why it matters:
The growing reliance on AI technology demands innovative solutions to the challenges posed by its identity requirements. As businesses continue to adapt and integrate AI agents into their workflows, managing their identities has become a pressing concern. Implementing IAM for AI ensures that sensitive data is protected from unauthorized access while promoting accountable and transparent interactions between humans and machines.
David Goldschlag’s Identiverse 2026 discussion brought much-needed clarity to this complex issue. His presentation was an insightful exploration into the real-world implications of IAM for AI and the essential elements required to successfully implement integrated identity frameworks that support both human-AI collaborations and large-scale data access requests. For businesses looking to ensure trust, stability, and growth in their AI-powered operations, understanding and integrating IAM solutions is a crucial step toward success.
Source: SC Media
