Maya Sherman is a senior advisor, defense technology researcher, and former Innovation Attaché at the Israeli Embassy in India. She consults globally with governments and think tanks on artificial intelligence, security, and emerging technologies. In this conversation, Sherman examines how AI is reshaping contemporary conflict—from the role of indigenous algorithms in the India-Pakistan standoff of Operation Sindoor to Israel’s rapid integration of AI systems post-October 7th. She also explores the dangers of digital disinformation, the rise of autonomous AI agents, and the evolving defense-tech partnership between India and Israel.
Que: I recently read an article you co-authored in The Jerusalem Post, titled “Two Battlefields, One War: One Year Since Operation Sindoor.” In it, you describe the conflict as a watershed moment for indigenous military AI, particularly in battlefield surveillance and electronic intelligence. How did integrating these systems transform India’s operational response and tactical decision-making?
Maya Sherman: Although Operation Sindoor was brief, it marked a fundamental evolution in how militaries conceptualize AI on the battlefield. While it was not the first instance of AI in combat, it was novel in how explicitly both sides framed the conflict as an “AI-driven” war.
This reflects a broader shift among scholars and state actors. Modern military powers no longer hesitate to acknowledge that automated, non-human assets provide a decisive tactical advantage. During the operation, AI dramatically enhanced real-time target accuracy and situational awareness.
Simultaneously, the conflict highlighted the double-edged nature of data flows. Despite its short duration, the sheer volume of information generated raised critical questions about how machine learning is used to execute—and counter—disinformation campaigns. Ultimately, the most transformative takeaway was that both militaries positioned AI not as a peripheral support tool, but as a core operational column.
Que: You mentioned information warfare. Your research highlights how digital propaganda campaigns persist long after kinetic engagements cease. How should defense establishments counter AI-generated disinformation that survives past a formal ceasefire?
Maya Sherman: Information warfare does not observe ceasefires. Because modern conflict is fundamentally tied to global data flows, the digital domain remains active indefinitely. We know from cyber espionage that digital environments obey entirely different strategic rules than physical battlefields.
To counter post-kinetic disinformation, states must build new defensive verticals. Militaries train extensively for physical combat, but civilians—who are far more exposed in the digital domain—receive almost no institutional training.
During Operation Sindoor, we witnessed sophisticated online campaigns circulating fabricated claims of military victories aimed directly at demoralizing the public. In high-stress scenarios, a population’s ability to discern truth from automated state propaganda becomes a matter of national security.
Technologically, AI functions as both the pathogen and the vaccine. While generative tools allow bad actors to fabricate and scale synthetic propaganda with precision, the same technologies can power real-time, decentralized fact-checking tools.
We must approach wartime information defense with the same gravity applied to election integrity—through representative multi-language datasets, rigorous algorithmic verification, and public digital literacy. Israel has had to learn this through continuous cycles of conflict: building civilian data resilience is essential to preventing foreign actors from weaponizing public anxiety.
Que: Turning to Israel’s military integration since October 7th—how has this high-intensity operational environment reshaped Israel’s approach to AI in intelligence analysis, surveillance, and tactical decision support?
Maya Sherman: Historically, Israel treats strategic crises as catalysts for technological pivots. Just as the COVID-19 pandemic forced the rapid adoption of automated diagnostic systems to address human bottlenecks, the operational shifts after October 7th accelerated defense-tech deployment at scale.When civilian populations are deliberately targeted—a reality Israel experienced acutely on October 7th and one India has faced repeatedly—the operational calculus changes. Protecting non-combatants in real time demands unprecedented accuracy and speed.
Israel’s strategic advantage has traditionally rested on cybersecurity and deep-tech control over the full domestic value chain. Recently, however, Israel established a dedicated national AI Directorate, channeling substantial capital into the ecosystem.
Beyond purely military applications, we are seeing a global surge in dual-use technologies—tools capable of shifting seamlessly between commercial and defense domains.
During the 12-day escalation with Iran in June, we observed intense state-sponsored digital campaigns aimed at narrative manipulation. The central challenge in countering online warfare is attribution; unlike physical battlefields, tracing an AI-driven narrative wave back to its sovereign originator in real time is remarkably complex. Crisis forces military and civilian infrastructure to operate as a single system, cementing AI as the foundational architecture connecting them.
Que: Information warfare often moves faster than ground combat. Looking at real-time data processing, air defense systems, and automated command structures, what core military lessons can be drawn from an AI framework—particularly following the escalation with Iran?
Maya Sherman: What is genuinely unprecedented today is how civilian data habits directly feed into the modern “fog of war.” Israeli soldiers undergo rigorous training, but civilians reacting to crises in real time are unequipped for the digital crossfire.
During missile strikes, panicked citizens uploading location footage inadvertently provided target acquisition data to adversary analytics engines.
AI’s primary strengths on the modern battlefield are clear: superior sensor fusion, automated asset tracking, faster target processing, and precise air defense interception. But the technology’s primary collateral damage is narrative clarity. Adversary state networks leverage generative AI to flood the domain with conflicting claims, making it difficult for the public—and analysts—to verify physical outcomes on the ground.
The real immediate risk isn’t an abstract “superintelligence taking over,” but the concrete reality of AI accelerating the fog of war, destabilizing civilian morale, and obscuring objective truth during critical operational windows.
Que: As AI tools compress military decision-making cycles from minutes to seconds, what are the primary strategic risks regarding automated escalation, lost human oversight, and algorithmic bias?
Maya Sherman: The conversation around autonomous AI agents is often distorted by corporate and political theater. While major tech conglomerates debate “algorithmic supremacy” and political leaders assert human control, we must remain critical about how modern warfare actually operates.
My primary long-term concern is not rogue code acting in isolation, but the convergence of robotics, autonomous software agents, and human-enhancement technologies—creating “hybrid” combat systems.
A far more immediate, concrete danger is democratized harm. Recently, reports emerged of non-state actors in Yemen using commercial, consumer-facing LLMs like Claude to assist in operational planning and strike logistics. Similarly, criminal actors use public models to learn weapon assembly.
Generative AI drastically lowers the technical barrier to entry. Extremist groups no longer require specialized software teams to execute sophisticated cyber actions or plan kinetic harm.
This places an unprecedented burden on commercial AI vendors and regulatory bodies to balance guardrails against over-censorship.
Furthermore, delegation of safeguard parameters to governments carries inherent risk. While democratic nations focus guardrails on counter-terrorism, authoritarian regimes routinely reframe civil rights or political dissent as “security threats” to justify algorithmic censorship.
Maintaining meaningful human control while preserving digital agency remains an unresolved global challenge.
Que: You previously served as Innovation Attaché at the Israeli Embassy in India. What structural bottlenecks exist in India’s quest to build sovereign military AI capabilities, and how can the India-Israel defense-tech partnership help bridge them?
Maya Sherman: The partnership between India and Israel is inherently complementary, anchored in asymmetric strengths.
Israel is an agile, deep-tech incubator. Due to our smaller geographic and demographic scale, we can coordinate across government, military, and private sectors rapidly, iterating advanced hardware and software internally end-to-end. However, Israel lacks scale.
India possesses extraordinary scale, a world-class engineering workforce, and a thriving commercial startup landscape. However, India’s challenges are vastly broader—spanning defense, climate, agriculture, and massive socio-economic diversity.
The primary structural bottleneck for India’s military AI ecosystem is the concentration of deep-tech capabilities within a few tier-one urban hubs. To build true national sovereignty in defense AI, India must bridge the technology gap between tier-one innovation centers and tier-two or tier-three regional ecosystems.
While India has successfully pioneered massive Digital Public Infrastructure (DPI) for civil society, the next milestone is developing specialized, sovereign deep-tech architectures. Israel cannot solve India’s vast scale requirements alone, but targeted joint ventures, co-development pilots, and deep-tech integration offer a clear roadmap for both nations to achieve military and technological independence.
