AI can sharpen foreign ministries' ability to manage interconnected crises but risks creating new vulnerabilities—from automation-driven entrenchment to an epistemic monoculture—so the authors urge 'augmented diplomacy' that keeps humans central to decision-making.

Main Finding

Polycrisis — simultaneous, causally entangled political, economic, environmental, technological, and social crises — raises the complexity of the foreign policy environment to a level that strains existing institutional and cognitive capacities. Artificial intelligence (AI) becomes an attractive corrective because it can amplify information collection, analysis, forecasting, and coordination across the foreign policy cycle. However, AI is ambivalent: it offers geopolitical advantage through improved complexity management but generates new strategic, institutional, and epistemic risks (notably a complexity paradox; complexity displacement; crisis-driven entrenchment; and the rise of an epistemic monoculture). The authors propose “augmented diplomacy” in which AI supplements — rather than replaces — human political judgment.

Key Points

• Definition and distinctiveness of polycrisis:
  • Polycrisis is an active, open-ended condition in which multiple real crises are causally entangled (shared stresses, domino effects, feedback loops), producing harms larger than isolated crises.
  • It differs from generic complexity, systemic risk, or VUCA by emphasizing simultaneity and entanglement across domains.
• Changing geopolitics:
  • Geopolitical advantage increasingly depends on the capacity to manage complexity, absorb shocks, and adapt — not just expansion or dominance. Resilience and adaptation coexist with traditional expansionist logic.
• Consequences for foreign policy:
  • Foreign ministries face growing inseparability of domestic and external issues and must coordinate across functional and regional silos under bounded rationality and resource constraints.
  • The policy problem shifts from addressing discrete threats to managing cross-domain interactions and emergent systemic consequences.
• Adaptation pathways:
  • Responses occur at institutional, individual, and technological levels. Institutional change is path-dependent and often incremental; crises create contested opportunities for reform.
  • Digital tools and AI are adopted as complexity-management instruments across information gathering, analysis, formulation, implementation, and evaluation stages.
• Benefits of AI:
  • Faster data collection, automated briefings, trend analysis, cost–benefit calculations, outcome forecasting, and improved whole-of-government coordination.
  • Use of AI can become a source of geopolitical advantage when it improves a state’s ability to manage polycrisis complexity.
• Risks and complications:
  • Complexity paradox: reliance on AI to manage complexity can produce new layers of complexity that are hard to control.
  • Complexity displacement: AI may shift rather than reduce complexity, offloading it to different actors, systems, or domains.
  • Crisis-driven entrenchment: rapid AI adoption during crises can lock in suboptimal architectures, vendors, or institutional dependencies.
  • Epistemic monoculture: widespread use of similar AI models and pipelines across actors increases correlated failures, manipulation risk, and reduced diversity of judgment.
• Normative prescription:
  • Augmented diplomacy: design AI systems to support human decision-makers, preserve human accountability, and avoid complete automation of political judgment.

Data & Methods

• Type: Conceptual analysis / theory synthesis.
• Methodology:
  • Integrates literature across geopolitics, foreign-policy analysis, institutional theory, digital governance, and AI ethics.
  • Develops a three-stage analytical framework: (1) establish the polycrisis condition and its burden on foreign policy; (2) trace institutional and individual adaptations and the technological response (role of AI across the policy cycle); (3) identify strategic, institutional, and epistemic complications introduced by AI.
• Empirical basis: No new primary data or quantitative testing — the contribution is synthetic and conceptual, mapping risks and opportunities and proposing a governance model (augmented diplomacy).

Implications for AI Economics

• Market demand and industrial structure:
  • Rising demand from states for AI systems tailored to complexity management (diplomatic analytics, forecasting, multilingual monitoring) will create specialized public-sector AI markets and consultancy services.
  • Procurement patterns may produce strong vendor lock-in and economies of scale for a few dominant suppliers (incentivizing oligopolistic rents and path-dependent lock-in).
• Public-good and coordination problems:
  • Information asymmetries, proprietary models, and security sensitivities create market failures (public-good aspects of shared situational awareness and the need for interoperable standards).
  • There is scope for government-sponsored open models or shared platforms to mitigate monopoly rents and reduce correlated vulnerabilities.
• Externalities and systemic risk:
  • Epistemic monoculture introduces positive correlation across actors’ beliefs and actions; common-mode failures create systemic tail risks (market-design externalities not priced by individual actors).
  • Economic models should internalize externalities from homogeneous AI use (e.g., through regulation, stress-testing, required model diversity, or insurance).
• Returns to AI deployment and cost–benefit metrics:
  • Need for new measures of the economic value of “complexity management” (reduced tail-risk exposure, faster response time, coordination gains) beyond conventional productivity metrics.
  • Evaluations must account for long-run costs of entrenchment, vendor lock-in, and loss of institutional diversity.
• Labor and skills complementarity:
  • AI is more likely to be complementary to high-level diplomatic judgment than a close substitute for it in complex political tasks; demand will shift toward workers with mixed technical and political skills.
  • Investment in human capital (training diplomats to use and audit AI) will become an important economic input.
• Strategic competition and uneven capacity:
  • Differential AI adoption creates asymmetric geopolitical advantages; countries with better AI-based complexity-management capabilities may lower the marginal cost of responding to polycrisis shocks.
  • This can widen global inequalities and produce second-order economic effects (trade, investment, and alliance patterns).
• Policy and regulatory economics:
  • Regulation to prevent epistemic monoculture, require human-in-the-loop, mandate model transparency/audits, and promote interoperable public infrastructure will influence market structure, R&D incentives, and cross-border data flows.
  • Subsidies or grants for public-interest models (open-source diplomatic AI) can correct under-provision and reduce correlated risk.
• Suggested economic research questions (operationalizable):
  • How to quantify the value of complexity management: metrics linking AI-enabled coordination to reductions in expected losses from cross-domain crises?
  • Modeling the probability and expected cost of common-mode failures produced by epistemic monoculture.
  • Welfare trade-offs between rapid AI adoption in crisis (short-term gains) versus long-term entrenchment and lock-in costs.
  • Optimal procurement mechanisms (auctions, modular procurement, open-source requirements) to balance innovation, security, and competition.
  • Returns to human–AI complements in diplomatic decision-making: empirical estimation of productivity gains when AI augments versus substitutes human expertise.
• Practical economic recommendations:
  • Favor modular, interoperable architectures and open standards to lower switching costs and limit vendor lock-in.
  • Invest in public or shared AI infrastructure for diplomacy to correct public-good under-provision and encourage model diversity.
  • Require stress-testing and red-team assessments that quantify correlated failure risks to inform procurement and insurance.
  • Subsidize training and human capital development to capture complementarities and reduce automation-driven displacement risks.
  • Implement procurement rules that emphasize vendor competition, model diversity, and auditability rather than lowest-cost turnkey solutions.

Short summary of limits: the paper is conceptual (no empirical testing), so economic implications outlined above should be followed up with empirical work to measure magnitudes, estimate costs of correlated failures, and assess welfare trade-offs.