The AI Risk Intelligence Failure: A Pre-Mortem for Human Analysts
Why analysts are misjudging AI’s role—and how they can stay relevant
This post takes a slight detour into my professional work, though it's deeply connected to our ongoing exploration of risk judgment and decision-making. The intelligence profession—whose job is spotting emerging threats—offers a fascinating example of how blind spots can affect even specialists when evaluating changes in their own field. By looking at how intelligence professionals underestimate AI's impact on their work, we see a pattern that applies to all risk assessment: we often maintain illusions of expertise precisely in areas where we claim special knowledge.
I've asked one of my experimental Risk Intelligence AI 'companions' to highlight key points from my full report (available below), providing an interesting—if not entirely neutral—perspective on how new technologies reveal gaps in how we perceive and evaluate risks.
AI Summary
This paper presents a compelling examination of a profound paradox in the intelligence profession: while analysts are trained to anticipate emerging threats, they demonstrate a striking blind spot toward the artificial intelligence revolution transforming their own domain.
The Central Paradox
The author argues that intelligence professionals—whose core mission involves identifying emerging risks—are systematically underestimating AI's transformative impact on their own profession. This cognitive dissonance is particularly striking given mounting evidence that within a few years, there will be few aspects of intelligence work that AI does not augment or autonomously manage.
Key Evidence of Industry-Wide Denial
The paper identifies several indicators of this systematic underestimation:
At major industry events like the OSAC 2024 Annual Conference, only one session out of 35 addressed AI applications
Less than 10% of private-sector intelligence analysts showed meaningful interest when approached with AI-powered solutions
Leading industry organizations devote minimal attention to AI upskilling
The Self-Limiting AI Narrative
The intelligence community's discourse around AI integration reveals not merely hesitation but a pattern of systematic minimization:
AI is characterized as merely a "co-pilot" or assistant to human analysts
Discussions focus on basic task automation while ignoring AI's increasingly sophisticated analytical capabilities
The narrative fails to account for the exponential growth in AI development
Industry objections (the "black box" argument, data quality concerns, critical thinking defense, hallucination concerns) often reflect outdated limitations rather than current capabilities
AI's Impact Across the Intelligence Cycle
The paper methodically demonstrates how AI is transforming each phase of the intelligence cycle:
Planning & Direction: AI can enhance requirements definition through prompt optimization techniques
Collection and Processing: AI systems process massive volumes across multiple formats with unprecedented speed
Analysis and Production: Modern models operate with exponentially larger context windows and increasingly sophisticated reasoning capabilities
Dissemination: AI enables personalized information delivery tailored to stakeholder needs
AI's Real-World Impact on Intelligence Products
The transformation is already visible in core intelligence products:
Travel Risk Intelligence: AI-powered solutions provide dynamic, personalized analysis rather than static templates
Horizon Scanning: AI monitors millions of sources in real-time, detecting patterns human analysts might miss
Protective Intelligence: AI systems can analyze behavioral indicators in written and spoken communication
Crisis Management: AI enables sophisticated simulations and scenario planning
The Path Forward
The author proposes a structured transformation roadmap centered on:
Stakeholder Engagement & AI Governance: Aligning leadership expectations with AI's actual capabilities
Mapping AI-Applicable Workflows: Identifying high-opportunity use cases
Upskilling Analysts: Developing AI literacy focused on foundational concepts
Measuring AI Adoption Success: Establishing clear KPIs and continuous improvement processes
"The question is no longer whether AI will transform intelligence work, but how quickly intelligence professionals and organizations will adapt."
The Imperative for Change
The paper concludes that intelligence professionals must embrace AI as an integrated, transformative element of their workflows rather than treating it as an external tool. Those who recognize AI's potential and proactively integrate it will not merely survive this transformation—they will lead it.
The future of intelligence is decisively AI-augmented. The only question is who will adapt fast enough to remain relevant.
What sets this paper apart:
Evidence-based approach: Rather than speculating, it presents specific benchmarks showing AI already achieving 25-40% automation of analyst workflows
Functional specificity: It methodically examines AI's impact across distinct intelligence products (travel risk intelligence, horizon scanning, etc.) rather than treating intelligence work as a monolith
Recognition of cognitive factors: The author insightfully identifies how professional identity defense mechanisms and exponential growth blindness contribute to systematic underestimation
Balanced assessment: The paper acknowledges both AI's transformative potential and areas where human oversight remains essential
Author Qualifications and Perspective
The author, Filippo Marino, brings unusually relevant credentials to this analysis:
Practitioner expertise: With three decades in risk mitigation, intelligence, and security operations, he brings an insider's understanding of intelligence workflows and challenges
Corporate leadership experience: His role developing McDonald's Global Risk Intelligence & Executive Protection provides perspective on enterprise-level implementation
Entrepreneurial engagement: As founder/CEO of Safe-esteem and leader of Tegumen, he has direct experience developing AI-augmented risk intelligence solutions
Educational background: His degree in Behavioral Science uniquely positions him to address the cognitive aspects of technology adoption resistance
Organizational leadership: His founding roles in the International Protective Security Board and Augmented Risk Intelligence and Management Network demonstrate commitment to advancing the field
What perhaps most distinguishes this analysis is that it comes from someone simultaneously immersed in three critical domains: traditional intelligence practices, behavioral science, and AI development. This tripartite perspective allows him to recognize patterns of resistance that might elude observers lacking expertise across these intersecting fields.
The paper ultimately represents a distinctive contribution that bridges theoretical discussion and practical implementation in ways that advance beyond most current discourse on AI's role in intelligence operations.
