Google early warning system fails to alert people during Türkiye earthquake
Google early warning system fails to alert people during Türkiye earthquake
Occurred: February 2023
Page published: October 2025
Google's earthquake early warning system failed to alert nearly 10 million people during Turkey's 2023 earthquakes, leaving millions without timely warnings and contributing to the many deaths and property damage and resulting in a public loss of confidence in technology-enabled alert systems.
Google's early warning system for earthquakes, specifically its Android Earthquake Alerts, failed to notify millions of people in Turkey during the country's 2023 earthquakes.
Despite the potential to alert around 10 million residents within proximity to the epicenter, only 469 "Take Action" alerts were sent out. The alerts are considered crucial for prompt safety measures.
Instead, a significant portion of the population received lower-level warnings or none at all, resulting in property damage, loss of life and other significant direct and indirect harms.
Over 55,000 people were killed and more than 100,000 were injured due to the twin earthquakes measuring 7.8 and 7.5 in magnitude that struck while many people were asleep in buildings that collapsed.
Google's earthquake alert system's detection algorithms underestimated the earthquake's actual magnitude, initially calculating it around 4.5-4.9 while the quake was a devastating 7.8 magnitude.
Consequently, only a tiny fraction of the roughly 10 million people within range received the crucial "Take Action" alerts that override phone settings to warn of imminent danger. About half a million others received a lower-level alert designed for mild shaking that does not override silent modes and was likely missed by most users, especially as the quake struck in the early hours while people were asleep.
This underestimation and the related algorithmic limitations meant the system failed to deliver life-saving warnings in time.
Google's system operates independently of national alert systems, and these shortcomings - coupled with a delayed and initially inaccurate public account of the system’s performance - highlight technical, transparency and accountability failures.
Google acknowledged these issues and revised its algorithms, but the incident spotlighted the intrinsic challenges in tuning early warning systems for large-scale natural disasters and the dangers of over-reliance on experimental technology without robust integration and transparency.
For those directly impacted, Google's failure contributed to a devastating loss of life and injury, and trauma.
More broadly, the incident exposed critical vulnerabilities in digital alert systems that millions rely on for safety during disasters, highlighted the ethical challenges of entrusting critical public safety functions to a private tech company, and contributed to a loss of trust in these systems, notably in Turkey.
It emphasised the need for robust, transparent, multi-layered early warning frameworks that serve as reliable supplements - not replacements - to national systems, ensuring vulnerable populations receive timely and actionable warnings to mitigate disaster impacts.
Developer: Google
Country: Turkey
Sector: Health; Politics
Purpose: Detect earthquakes
Technology: Machine learning
Issue: Accountability; Accuracy/reliability; Transparency
AIAAIC Repository ID: AIAAIC2082