The 8.8 Wake-Up Call: Emergency Communication Lessons from the Pacific Rim

Dear {{ first_name | reader }},

The phone didn't ring at 3 AM this time. Instead, seismographs around the Pacific Rim lit up like Christmas trees as an 8.8 magnitude earthquake struck off Russia's Kamchatka Peninsula on July 30th.

Within minutes, tsunami warnings cascaded across the Pacific from Japan to Chile and Hawaii to Ecuador. What followed was emergency communication on a scale rarely seen. And it worked.

In this week’s edition of Wag The Dog, I take a closer look at the risk and emergency communication that took place during the earthquake and tsunami.

Looking forward to reading your feedback.

Picture this: a submarine earthquake 20 kilometres deep generates waves up to five metres high. The Pacific Tsunami Warning Center has minutes to alert millions of people across multiple continents and time zones.

Japan evacuates nearly 2 million people. Hawaii closes beaches. Chile empties coastal towns. Russia moves entire communities inland.

The outcome? Even though the earthquake was very strong and caused a large tsunami, no one died directly from the tsunami. This wasn't just luck; it shows how effective emergency communication can be.

Speed Made the Difference

Ten minutes. That's how long the Pacific Tsunami Warning Center took to issue initial alerts after the earthquake struck. In that brief window, they triggered a communication cascade that reached millions of people hours before the waves arrived.

Japan's response deserves special attention. The country's J-Alert system, combined with drone-deployed warnings in coastal areas, shows how far emergency communication has evolved since the devastating 2004 Indian Ocean tsunami. When 2 million people evacuate in an orderly fashion, that's years of public education paying off.

Technology Worked Under Pressure

International cooperation happened at breakneck speed. The Pacific Tsunami Warning Center, the Japan Meteorological Agency, NOAA, and national warning centers across multiple continents shared information seamlessly. Modern warning systems integrated data from seismometers, DART buoys, and coastal tide gauges to provide real-time wave height forecasts and precise arrival predictions.

Hawaii's experience shows how far we've come. The state's comprehensive alert system and well-marked evacuation zones demonstrate years of preparation paying off.

AI Gone Wrong

Perhaps most concerning was the role of artificial intelligence in spreading dangerous misinformation during the crisis. This incident has become a stark example of what happens when untested AI systems are deployed in real-time emergency situations.

Grok AI on X (formerly Twitter) created serious public safety risks by falsely claiming Hawaii's tsunami warning had been cancelled hours before authorities officially lifted it¹.

Multiple eyewitnesses and technology media documented that Grok responded to user queries with messages indicating the warning was over at 3:23 p.m. local time, when in fact the official warning remained in place for many more hours.

Social media posts with screenshots captured these errors, and the blunder drew widespread criticism for potentially endangering public safety.

Google's AI-generated Overviews proved equally problematic. Reports indicate that Google's AI summaries - which often appear prominently atop search results - provided misleading or "dangerously wrong" advice about evacuation protocols and the official status of alerts in Hawaii and possibly elsewhere.

These AI outputs failed to update reliably with verified information as the situation developed, creating confusion precisely when people needed accurate, up-to-date guidance.

The broader implications are deeply troubling. As AI-based information tools become increasingly embedded in online platforms and search engines, their inaccuracies and "hallucinated" (fabricated) outputs during emergencies can have real-world consequences.

Technology, disaster response, and risk communication analysts have pointed out that this incident highlights the risks of deploying large-scale, untested AI systems without proper constraints and supervision during high-stakes information delivery tasks.

Both Grok and Google's AI systems failed when it mattered most, amplifying uncertainty and confusion during a disaster rather than providing the reliable information people desperately needed. This reveals a critical vulnerability in our increasingly AI-dependent information ecosystem.

The Uncertainty Challenge

Tsunami warnings lasted many hours as waves crossed the Pacific. While extensive warnings ultimately proved more cautious than necessary, authorities struggled with communicating the inherent uncertainty in tsunami predictions.

How do you explain why massive evacuations were necessary for waves that ultimately proved smaller than feared? This challenge creates public scepticism and potential future compliance issues.

Invest Before You Need It

The relatively smooth evacuations in Japan reflect years of public education about tsunami risks and evacuation procedures. Hawaii has similarly invested in extensive tsunami zone signage so residents know where danger zones are and how to evacuate quickly.

Your organization needs this same foundation. Emergency communication doesn't start when the emergency hits. It starts with the groundwork you lay today.²

Monitor and Counter AI Misinformation

The Grok and Google AI failures during this tsunami response reveal a new challenge for emergency communicators: AI systems can spread dangerous misinformation faster than you can correct it. Your crisis communication plan must now include monitoring major AI platforms for false information about your emergency.

Establish relationships with platform providers before a crisis hits, so you have direct channels to flag and correct AI-generated misinformation quickly. More importantly, invest in building your organization's credibility as the authoritative source for information.

When AI systems fail, people need to know where to find reliable updates. Make sure your official channels are easily identifiable, frequently updated, and consistently referenced in all your communications.

The goal isn't to prevent AI misinformation entirely (that's beyond your control) but to ensure people know how to find and verify accurate information when confusion spreads.

Clarity Trumps Cleverness

Warning communications used unambiguous language: "Take Action! Destructive tsunami waves expected" and "Evacuate immediately from coastal and riverside areas." No corporate speak. No hedging. Just clear, actionable direction.

When your crisis hits, your audience won't have time to decode your message. Make it crystal clear what they need to do.

Plan for Communication Failure

Remote Russian communities experienced power outages and mobile service failures immediately following the earthquake. This compromised ongoing emergency communications precisely when they were most needed.

Your emergency communication plan must account for infrastructure failure. What happens when your primary communication channels go dark?

This tsunami response shows international emergency communication at its best. The coordination prevented mass casualties through rapid response. Yet it also revealed vulnerabilities that need ongoing attention.

The disruption of AI systems providing false information during the crisis reminds us that our digital communication landscape brings both opportunities and risks. As these technologies become more integrated into our systems, we need to manage their potential for both help and harm.

The earth will shake again—metaphorically if not literally. Product recalls. Data breaches. Executive scandals. Workplace accidents. The specific crisis doesn't matter. What matters is whether you've built the communication infrastructure to respond effectively.

The Pacific tsunami response offers a blueprint: invest in pre-event education, use clear and actionable messaging, plan for multi-modal communication, prepare for infrastructure failure, and coordinate across all relevant stakeholders.

Most importantly, remember that speed and clarity can literally save lives, or at minimum, save your organization from becoming another cautionary tale.

What aspects of this tsunami response could your organization learn from? I'd love to hear your thoughts. Just reply to this email and let's discuss how these lessons apply to your specific context.

Your AI is only as good as the data you feed it. If you're building or fine-tuning generative models, Shutterstock offers enterprise-grade training data across images, video, 3D, audio, and templates—all rights-cleared and enriched with 20+ years of human-reviewed metadata.

With 600M+ assets and scalable licensing, our datasets help leading AI teams accelerate development, simplify procurement, and boost model performance—safely and efficiently.

Book a 30-minute discovery call to explore how our multimodal catalog supports smarter model training. Qualified decision-makers will receive a $100 Amazon gift card.

Book a call

_{For complete terms and conditions, see the offer page.}

PS: I hope you've enjoyed this newsletter! Creating it each weekend is a labour of love that I provide for free. If you've found my writing valuable, the best way to support it is by sharing it with others. Thank you for reading!

_{Parts of this newsletter were created using AI technology to draft content. In addition, all AI-generated images include a caption stating, 'This image was created using AI'. These changes were made in line with the transparency requirements of the EU AI law for AI-generated content. Some links in this newsletter may be affiliate links, meaning I earn a small commission if you click and make a purchase; however, I only promote tools and services that I have tested, use myself, or am convinced will make a positive difference.}