Original broadcast 10/7/25
Presented by Semantic Visions & Carahsoft
The intelligence community is facing a future defined by emerging risks that are more complex, fast-moving, and disruptive than ever before. Among these challenges are synthetic media, the convergence of state and criminal actors, and vulnerabilities in global supply chains. Each of these threats has the potential to destabilize information environments, disrupt economies, and undermine national security.
Sears identified synthetic media as perhaps the most immediate challenge. The rise of generative artificial intelligence has enabled adversaries to produce manipulated narratives, deepfakes, and fabricated content at unprecedented speed and scale. “Automated narrative signals can move crowds within hours,” he warned. Unlike traditional disinformation campaigns that unfold over days or weeks, synthetic media can reshape public perception in real time. This speed makes it far harder for intelligence organizations to respond before damage is done.
The second risk is the growing convergence between state actors and criminal groups. Nation-states are increasingly leveraging cybercriminal syndicates, cartels, and other illicit networks to conduct operations that advance political objectives while masking attribution. This blending of motivations—profit for criminals, influence for states—creates hybrid threats that defy traditional categories. Intelligence agencies must now contend with adversaries that use the tools of organized crime to execute state-sponsored objectives, complicating both detection and response.
The third area Sears highlighted is supply chain vulnerability. Critical industries, from energy to chemicals to advanced manufacturing, rely on complex global supply chains that are increasingly exposed to disruption. He noted that early warning signs of these risks often surface first in open-source environments, particularly in fringe or foreign-language media outlets. Detecting and analyzing these signals before they cascade into larger crises requires robust OSINT capabilities, including multilingual coverage.
Sears explained that the intelligence community’s greatest challenge is not the lack of data, but rather the overwhelming volume of it. For years, analysts struggled with what he called the “too much problem”—too much content to monitor, too many platforms to scan, too many languages to translate. Even in English alone, the scale of information was daunting. When multiplied across dozens of languages and hundreds of outlets, the challenge became nearly unmanageable.
He noted that many intelligence organizations still rely on traditional Boolean string searches in a single language, which leaves them blind to critical signals. Newer models can break queries into semantic components, translating them across languages and curating the results. This approach reduces noise and provides structured outputs that highlight actionable narratives rather than raw, unfiltered data. As a result, even relatively junior analysts can produce insights that once required years of expertise.
Still, technology alone cannot solve the problem. Synthetic media, for example, presents a constant cat-and-mouse game. As AI-generated disinformation becomes more sophisticated, defenders must train their own models to detect and counter it. Sears cautioned that the battle will increasingly be “your model versus my model,” with adversaries and defenders racing to outpace each other. In such an environment, human judgment remains vital to assess provenance, credibility, and context.
Another dimension is the need for explainability and provenance in OSINT tools. Many chat-based AI systems can generate answers but offer little transparency into how those answers were derived. For intelligence work, that is unacceptable. Analysts need confidence that their conclusions are based on verifiable sources. Sears emphasized that solutions must include mechanisms to trace findings back to their origins, ensuring accountability and accuracy in decision-making.
Looking ahead, he predicted that the biggest challenge in the next year will be integrating multiple layers of data—classified and unclassified—into coherent, shareable pictures. OSINT is supposed to accelerate targeting and reduce decision cycles, but integrating open-source insights with classified intelligence remains difficult. Building pipelines that allow simultaneous analysis of both will be crucial for coalition operations, where partners often cannot access U.S. classified material but can share unclassified assessments derived from open sources.
The risks Sears outlined—synthetic media, state-criminal convergence, and supply chain fragility—are not hypothetical. They are already manifesting today. But their scale and intensity will only increase in the years ahead. The intelligence community must adopt new technologies, expand multilingual monitoring, and build resilient data pipelines to stay ahead of these threats.
As the line between information and disinformation blurs, and as criminal and state actors collaborate in new ways, speed and accuracy will determine who gains the upper hand. For intelligence professionals, preparing for this future means investing not only in advanced tools but also in the training, tradecraft, and judgment required to navigate an increasingly chaotic information environment.
Key Takeaways
Synthetic media powered by generative AI can alter public perception in hours, creating urgent challenges for intelligence agencies.
The convergence of state and criminal actors creates hybrid threats that are harder to attribute and counter.
Multilingual AI models are transforming OSINT workflows, but human judgment and provenance remain essential for accuracy.