Today, foreign ministries across the world are drowning in information — news reports, intelligence assessments, social media posts, satellite imagery, economic data, speeches and diplomatic cables.
AI can summarize thousands of documents in minutes, track political sentiment across countries, detect emerging crises earlier, analyze sanctions, dissect trade flows and monitor military movements.
The technology’s promise is compelling. A foreign minister who once relied on a 20-page briefing note may soon receive an AI-generated dashboard constantly updated with developments from across the globe.
Negotiations could become more data-driven. AI can model trade agreements, energy partnerships, climate commitments and infrastructure projects, among others.
Before entering negotiations, governments may simulate multiple scenarios and identify optimal bargaining positions. For example, while negotiating a free trade agreement, diplomats could rapidly assess the impact of thousands of tariff combinations.
Traditionally, major powers held an advantage because they could afford large diplomatic corps and intelligence agencies. But AI could partially level the playing field for smaller nations with fewer diplomats.
With access to advanced AI tools, smaller nations may be able to perform analytical tasks previously requiring hundreds of specialists. This could democratize diplomatic capability, though access to the most advanced AI systems may itself become a source and determinant of power.
Yet diplomacy has an enduring complication: A single leader can upend all predictions with one unexpected decision.
Donald Trump’s tariff escalations and policy reversals, Kim Jong Un’s unexpected military alignment with Russia, and Benjamin Netanyahu’s carrot-and-stick approach in Gaza are stark examples. The Iranians walking out of talks after Trump’s recent threat is another. Can diplomacy really be automated when leaders themselves are so often unpredictable?
Political leaders are not algorithms. The central assumption behind most predictive systems is that actors behave according to identifiable patterns. AI learns from past behavior, detects correlations and estimates probabilities.
This works remarkably well when patterns are stable. Diplomacy, however, often turns on individuals who deliberately refuse to behave predictably — or, in some cases, rationally. This is where the limits of algorithmic diplomacy become starkly evident.
Consider Trump. During his second presidency, many analysts anticipated tougher trade measures against China. What proved difficult to anticipate was the speed, scope and sequencing of those decisions. Markets, businesses and even allies found themselves constantly revising expectations as Trump raised and lowered tariffs at will.
For Trump, unpredictability is a strength, not a weakness, and is often leveraged as a negotiating tool. Threatening extreme measures, altering positions, creating uncertainty and keeping counterparts guessing can all generate leverage at the bargaining table. The more others try to predict his actions, the greater the advantage of remaining unpredictable.
A similar dynamic can be observed in North Korea. For years, many assumed Kim Jong Un remained heavily dependent on China and therefore constrained by Beijing’s preferences.
Instead, he deepened military cooperation with Russia, extracted strategic and economic benefits from that relationship and simultaneously strengthened his bargaining position with China.
Kim has also repeatedly alternated between threats and conciliatory gestures, making it difficult for outside observers to determine whether particular signals represent genuine policy shifts or tactical maneuvers.
Benjamin Netanyahu presents a different challenge. Throughout the Gaza conflict and wider regional tensions, outside observers often struggled to predict Israeli decision-making.
Military considerations were only one factor. Coalition politics, domestic pressures, personal political survival and strategic calculations all interacted in complex ways.
Such variables are difficult to fully capture in datasets or predictive models. None of these examples prove that world leaders have “defeated” artificial intelligence — governments do not publicly disclose what their predictive systems forecast.
But they do illustrate a broader point: The same leaders who surprise intelligence agencies, diplomats, markets and policy experts are also likely to expose the limitations of AI-driven prediction.
The problem is not that AI lacks intelligence. The problem is that politics is not governed solely by logic. Diplomacy involves trust, prestige, fear, ambition, ideology, ego and perception.
Leaders sometimes pursue courses of action that appear costly or irrational to outside observers because they are responding to domestic political pressures, personal convictions or concerns about reputation. In many cases, these motivations are difficult to quantify and even harder to predict.
Paradoxically, the growing use of AI may actually increase the value of unpredictability. If governments increasingly rely on predictive algorithms, leaders gain incentives to become less predictable.
Strategic ambiguity becomes a source of power. Sending contradictory signals, changing course unexpectedly or maintaining uncertainty can complicate an adversary’s calculations and reduce the effectiveness of predictive tools.
History offers an instructive precedent. During the Cold War, President Richard Nixon’s “madman theory” sought to convince adversaries that he might act unpredictably, thereby encouraging caution. Whether the strategy succeeded remains debated. What matters is the underlying logic: uncertainty can itself be a strategic asset.
The leaders most likely to frustrate predictive systems are not necessarily irrational. They are often leaders who understand that others are trying to predict them and deliberately cultivate ambiguity.
In a world increasingly dependent on algorithms, strategic unpredictability may become even more valuable. Leaders like Trump consciously exploit AI’s weaknesses: appearing unpredictable becomes strategically valuable when governments rely on predictive algorithms.
A leader who frequently changes positions, sends contradictory signals or makes unexpected decisions becomes harder for AI systems to model — and in that sense, erratic behavior can itself become a negotiating tool.
This does not mean AI will fail in diplomacy — far from it. AI will transform diplomatic practice in profound ways: improving intelligence analysis, accelerating decision-making, enhancing crisis monitoring and strengthening policy planning. It may even help smaller countries compensate for limited diplomatic resources.
For India, these advantages are particularly significant. India’s diplomatic service remains relatively small for a country of its size and global ambitions. AI could serve as a force multiplier, improving analytical capacity and enabling more effective engagement across multiple regions simultaneously.
As AI becomes increasingly central to national power, investments in indigenous technological capabilities, computing infrastructure and advanced research will become essential components of strategic autonomy.
Yet even the most sophisticated AI system cannot fully answer the questions that matter most in diplomacy. Is a leader signaling resolve or bluffing? Is a threat genuine or intended for domestic audiences? Is a sudden conciliatory gesture the beginning of a compromise or merely a tactical pause?
Such questions are not fundamentally data problems — they are problems of interpretation and judgment. And as AI handles more routine analysis, human diplomats may become more valuable, not less.
A seasoned ambassador might notice a leader’s mood, body language, personal insecurities, rivalries within the ruling elite or changes in tone during private conversations.
These subtle signals often matter more than data. An AI might conclude: “There is a 75% probability of agreement.” A diplomat might say: “The leader feels personally insulted — the deal is unlikely to survive.” The second assessment could prove more accurate.
That is why diplomacy is unlikely to become AI-led and far more likely to become AI-assisted. Algorithms will process information faster and more comprehensively than any human bureaucracy, identifying patterns, generating scenarios and highlighting risks.
But the most consequential decisions in international politics will continue to be made by human beings — leaders who can choose, at any moment, to follow the script, rewrite it or discard it altogether. And that is precisely why the leaders most likely to succeed in advancing national interests may be those who can still break the algorithm.
Raghu Gururaj is a retired Indian ambassador and former foreign service officer.
