Multiplayer AI: The New Operating Model For Identity Security

Forbes

3 hours ago

Dr. John Pritchard is the Chief Product Officer at Radiant Logic, responsible for the company's global product vision.
AI-powered deepfakes and credential attacks are rewriting the rules of cyber risk, with identity-related breaches now costing organizations an average of $4.45 million per incident and accounting for over 70% of successful attacks on enterprise infrastructure and supply chains. Despite record investments in detection and response, breaches keep making headlines.
Why?
I call this the identity security paradox: More technology doesn't equal protection, especially if tools—and the people and AI agents using them—don't work together.
Identity is the primary attack surface in the enterprise. Most organizations built their identity security stack on a traditional combination of IAM, IGA and PAM, but the rapid proliferation of cloud apps, machine identities and AI agents outpace these traditional controls. The result? Siloed data, unmanaged privileged accounts and hidden nonhuman identities—each a potential attack vector.
Gartner finds that 65% of organizations still lack IAM maturity, weighed down by technical debt and fragmented architectures. Point solutions deployed to 'fix' audit findings or compliance gaps create more complexity, not less. Attackers exploit these seams, moving laterally between systems and identities that aren't monitored holistically.
CISA's Silentshield Red Team Assessment demonstrated that decentralized teams and poor communication allowed adversaries to persist undetected, even when individual groups spotted anomalies. The lesson is clear: Solo efforts—whether a lone expert, an isolated AI agent or a disconnected tool—cannot keep pace with adversaries who are increasingly agile, automated and collaborative.
To close these gaps, interoperability must become the standard for tools and the people and AI agents using them. Interoperability means more than connecting dashboards or sharing alerts. It's about ensuring that identity security posture management (ISPM) and identity threat detection and response (ITDR) systems share data, context and workflows in real time, across both human and machine identities.
Gartner recommends a 'system of systems' approach, built on identity fabric principles, to support zero trust and intelligent automation. This means breaking down technical and organizational silos so prevention and detection teams operate from a unified, continuously updated single source of truth for identity data—a concept Gartner identifies as foundational for modern identity security. This trusted, authoritative data layer enables faster, more accurate decisions and ensures that every team acts on the same intelligence. When ISPM and ITDR interoperate, and when human and AI teammates collaborate based on shared reference points, blind spots shrink and attackers have fewer seams to exploit.
I call the next evolution in identity security: multiplayer AI—intelligent systems designed to amplify human capabilities through enhanced teamwork. Gartner predicts by 2027, 90% of successful AI implementations in cybersecurity will focus on tactical task automation and process augmentation, not full autonomy or staff replacement.
Multiplayer AI enables human and AI collaboration, breaking down silos and bridging gaps between prevention and detection. AI excels at analyzing vast datasets, detecting patterns humans miss and automating repetitive processes. Critical decisions, like determining whether anomalies are a threat or false positives, still require human judgment and contextual understanding. Studies show organizations using collaborative AI models—human decisions based on AI recommendations—see faster response times, fewer security incidents and improved resilience.
The key is not just technology, but teamwork: AI handles the heavy lift of data processing and pattern recognition, while humans provide creativity, ethical oversight and business context.
Thankfully, the industry is moving quickly. With the meteoric rise of agentic AI, open standards like Model Context Protocol (MCP) and Agent2Agent (A2A) are enabling AI agents from different vendors, clouds and frameworks to communicate, share context and coordinate tasks securely. Technology partners including OpenAI, Microsoft and Google are already adopting these protocols, breaking down silos that limit automation's impact.
For business leaders, agentic AI means specialized agents for threat detection, access management, compliance and user behavior analytics can now form ad hoc teams-automating complex workflows and adapt to new threats. By 2028, Gartner forecasts multiagent AI will account for 70% of threat detection and response implementations, primarily to augment—not replace—staff. Early adopters will see measurable results: Leveraging agent-to-agent collaboration is predicted to cut attacker dwell time in compromised environments by up to 50%, while accelerating response and reducing operational risk. When humans and AI work together, identity security becomes faster, smarter and more resilient.
1. Establish an interoperability baseline. Audit ISPM and ITDR tools for data sharing and workflow integration across human and machine identities. Ensure architectures support agent-to-agent interoperability using open standards like MCP and A2A, so specialized agents can collaborate and automate cross-vendor workflows. Set quarterly targets to reduce IAM tool integration gaps.
2. Pilot tactical AI augmentation. Start with a focused, data-driven use case, such as automated privilege review or anomaly detection. Track improvement in response time and risk reduction.
3. Build AI literacy and human oversight. Train teams on both the benefits and limits of AI, including where human verification is required in critical workflows.
4. Continuously review identity hygiene. Use AI-driven discovery to identify unused or risky accounts, but require human validation before making changes. Aim to reduce privileged account sprawl and remediate orphaned accounts as they are detected.
5. Measure what matters. Track outcome-driven metrics such as percentage reduction in excessive permissions, improvement in MFA deployment rates and decreased incident response times. For example, reducing excessive permissions by 20% and increasing MFA coverage to 95% of privileged accounts within one year.
The next breach won't be stopped by just another dashboard or a new AI agent. Organizations that have achieved true interoperability across tools, teams and AI will be able to respond more accurately to security issues. Multiplayer AI and agent-to-agent collaboration will lead the blueprints for resilience in the age of AI turbulence.
Start by assessing your current environment for interoperability gaps, unify your tools and teams and empower your people with AI that amplifies—not replaces—their expertise. In the high-stakes game of identity security, victory belongs to those who play as a team.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?