Did We See It Coming? What a Decade of Crises Teaches Us About Prediction

A global pandemic, multiple wars, and the rise of AI are just some of the crises and disruptions the last decade has witnessed. Did we see it coming? At a special webinar marking the 10th anniversary of NYU's Riskathon, the simulation’s lead analyst, Professor Maha Hosain Aziz, revisited predictions first made in 2016 and measured them against today's fractured geopolitical landscape. Her conclusion: many of today's most destabilizing crises weren't sudden shocks. They were slow-burning trends we watched unfold in real time

Dr. Maha Hosain Aziz is a professor in NYU's MA International Relations program, specializing in global risk and prediction, a World Economic Forum Global Future Council expert, and award-winning author of Future World Order and The Global Kid. She serves as lead analyst for the NYU Riskathon, a decade-long global risk prediction simulation she runs with Wikistrat and her graduate students.

Watch full recording:

Drawing on a decade of crowdsourced simulations, Prof. Aziz revealed both the power and limits of forecasting the future. Here's what ten years of data taught us.

1. America's Declining Legitimacy: Trend, Not Blip

Back in 2016, early Riskathon simulations asked whether America's unipolar moment was ending. A decade of data later, the answer is clear: US legitimacy has declined consistently, year after year. But rather than witnessing a clean transition to multipolarity, Prof. Aziz argues we're entering something messier: a polycentric order where smaller states, non-state actors, and technology firms wield outsized influence while major powers struggle with domestic burdens and strategic overreach.

2. Democratic Backsliding Is Structural, Not Cyclical

The crisis of political legitimacy identified in 2016 has only intensified. Democratic decline now spans roughly two decades, with protests, polarization, and institutional distrust recurring across every region. Social media and AI have accelerated this erosion, amplifying challenges to authority and raising an uncomfortable question: are we experiencing a temporary downturn or entering a post-democratic era?

3. Economic Nationalism Became the New Default

While specific shocks like tariff wars and inflation-driven unrest weren't fully anticipated, simulations correctly identified the underlying trajectory. The crisis of economic legitimacy, driven by inequality, youth unemployment, and populist backlash, was visible years ago. Now AI adds a volatile new layer, intensifying anxiety about jobs, economic purpose, and social stability, particularly in advanced economies.

4. Identity-Driven Polarization: The Most Under-Addressed Risk

Among all risk themes tracked since 2016, Prof. Aziz emphasized that identity remains the least effectively managed by policymakers. Anti-minority sentiment, xenophobia, and extremism haven't receded. They've evolved, increasingly amplified by digital platforms and AI tools. Her assessment is stark: without deliberate intervention, this trend will worsen, not stabilize.

5. Shock Events Reveal Both Foresight and Blind Spots

Recent years introduced shock scenarios designed to surface low-probability, high-impact disruptions. Some have already materialized, including political reversals once deemed unlikely. Others remain worryingly plausible: private actors pursuing radical climate interventions, the resurgence of environmentally motivated extremism, and long-term challenges to dollar dominance through alternative currencies and digital finance.

6. AI's Most Dangerous Risks Are Social, Not Technical

The most underestimated AI risks aren't about runaway algorithms or job displacement numbers, Prof. Aziz warned. They're about mental health pressures, occupational identity loss, and a slower-than-expected transition to new forms of work. These social disruptions could fuel wider instability through 2030 and demand as much attention as productivity gains or national competitiveness.

7. Crowdsourcing Works: The Case for Collective Intelligence

Prof. Aziz also reflected on methodology, offering a defense of crowdsourced risk analysis.  Its core strengths, she argued, lie in empirical richness and the diversity of perspectives it brings together. While she acknowledged inherent biases and blind spots, she framed these as opportunities for improvement rather than fatal flaws, suggesting AI could help audit forecasts and identify gaps. Her central point: collective intelligence offers our best path forward precisely because it captures what isolated experts might miss.