Identity and Access Management: The Next Frontier with AI

When developing AI-based capabilities for Identity and Access Management (IAM) systems, organisations must often choose between improving user productivity and enhancing security measures. Some Artificial Intelligence (AI) capabilities streamline user operations and improve efficiency, while others aim to tighten security.

These features help customers work more efficiently and shorten the learning curve. Some AI-based features, like chatbots, focus on productivity, while others, like policy recommendation engines, combine productivity with security. These features can streamline users’ work and provide collective or heuristic-based knowledge, guiding them to make better decisions and enhance security. Other capabilities, such as discovering and alerting suspicious activities, lean even further toward the security end of the spectrum. 

The Three Pillars of AI in IAM 

Looking into the future, we can expect a wide range of technologies to emerge in each category, from the near and achievable to the more distant and complex. AI in IAM can be categorised into three main pillars, each blending productivity enhancements with security improvements. The below illustrates better how we can build these pillars.

1) Chatbots and AI Assistants

AI-driven chatbots and assistants in IAM will soon offer more than text-based Q&A interactions. They will provide context-aware recommendations and integrate with third-party systems. These AI tools will understand users' needs and unique customer circumstances, making operations more intuitive, efficient, and tailored. By answering queries, suggesting next steps, and executing commands, AI assistants will become indispensable in the IAM toolkit.

Here’s what’s likely to roll out in the next few years in this area of AI in IAM:

• Documentation chatbots will continue to evolve and improve their ability to answer generic questions based on the body of documents, knowledge base and other sources of information.
• Assistant chatbots will understand natural language and execute complex commands. For example, user queries may involve a chain of API calls, requiring a precise understanding of the parameters for each call and delivering the appropriate response.
• Context-aware chatbots and assistants will be more knowledgeable about individual user circumstances than their current AI predecessors. Rather than providing the same outputs to different users, these next-gen chatbots will “know” things about you and tailor their responses accordingly.
• Automatic issue detection and guided remediation will enable AI chatbots and assistants to become more proactive, suggesting actions and next steps. Whether it’s the next item on your to-do list or resolving an error on your screen, these assistants will increasingly offer solutions to problems and tasks – asking only for your confirmation.

2) Access Policies

AI technology will transform access policy definitions by generating dynamic least privilege policies based on natural language and intent. IAM administrators will become strategic supervisors, setting guidelines, approving policies, and managing anomalies. This shift will speed up tasks and reduce the need for technical knowledge. While it raises concerns about precision, AI-generated policies are likely to enhance security by reducing human error and misconfiguration.

Here’s what to look for in the next few years:

• Policy recommendations based on best practices, or the collective knowledge of other customers is a trend consumers can soon expect.
• Intent as policy might be one of the most revolutionary AI promises to the world of IAM. We’ll see natural language used to define new and explain existing policies. The intent will become the policy. Automatic policy creation is the logical next step after policy recommendation. Such policies would rely on history or heuristics for their creation.

3) Risk-based Access

The third pillar focuses on the nature of access itself. As AI makes access to systems more personalised and contextualised, access becomes more dynamic and transparent. This means fewer repetitive logins and multi-factor authentication (MFA) prompts during normal operations, leading to smoother workflows and less user frustration.

Here’s what we may see in this area in the near and not-as-near future:

• Activity summaries and security insights will be generated from the user’s interactions with systems, which produce a digital trace. Generative AI (GenAI) will transcribe and summarise this trace into human-readable text. Additionally, GenAI will alert you if you perform a risky operation during the session.
• Behavioural profiling and threat detection will work together to create and continually update risk profiles for workloads and users. As a result, more granular and precise risk-level management and threat detection will be achievable.
• Automated threat prevention will be the next natural step following the arrival of threat detection mechanisms. It will likely take many forms, such as stopping a suspicious session or suspending or requiring additional login measures for a questionable user.
• Automatic policy creation (extension): With the ability to maintain user-specific profiles, systems will use this data to create user-specific and context-specific behaviour, resulting in more personalised and dynamic access policiesIntegrating AI in IAM is an ongoing journey toward creating more secure, efficient and user-friendly systems. As we look to the future, the focus will be on how AI can seamlessly integrate into the core areas of IAM to provide increased security and productivity.

Written By - Rohan Vaidya, Area Vice President, India & SAARC, CyberArk

Read More:

Channel Plays a Critical Role for AI-driven Cybersecurity Solutions