Over the past year, a bizarre rumour has been circulating online: ask Google Assistant or Amazon’s Alexa when World War III will begin, and they’ll confidently give you a date, a time, and even a trigger event. The answer — “November 23, 2023, at 6:05 p.m., when Russia attacks Germany” — has spread across TikTok, YouTube, and other platforms, sparking everything from nervous jokes to outright conspiracy theories.
How the rumour spread
The question “When will World War III start?” has gained more attention since Russia’s invasion of Ukraine. Political leaders, including Emmanuel Macron, have addressed the risk of escalation, so it’s no surprise the topic has seeped into public curiosity — and onto our smart speakers.
Videos of Alexa and Google Assistant delivering the same chilling “prediction” have gone viral, with users filming their interactions. Some clips date back to 2021, long before the current conflict, suggesting this wasn’t a new glitch. But in France, recent global events amplified the fascination, and in true internet fashion, misinformation followed quickly behind.
Why smart assistants can be confidently wrong
Voice assistants like Google Assistant and Alexa are designed to provide quick answers to almost any question, pulling from the top-ranked or most relevant sources in their database. The problem? Those “sources” can include unreliable information — and the algorithms don’t always distinguish between fact and fiction.
In this case, Alexa’s fabricated answer had become the top result for the query. Google Assistant in France simply echoed it, unintentionally passing along the false claim. The result was an AI-assisted game of telephone, where a made-up “fact” was repeated until it sounded authoritative.
Google changes course
More recently, Google Assistant has updated its reply. Now, instead of parroting Alexa’s words, it redirects users to Wikipedia, where World War III is described as a hypothetical global conflict, often imagined in the context of nuclear warfare or other large-scale hostilities. This more cautious approach avoids making a concrete — and completely unfounded — prediction.
A broader effort to curb misinformation
In August, Google confirmed it had reviewed how it answers certain “odd or speculative” questions. The move is part of a wider attempt to combat misinformation and avoid giving definitive answers where no verified information exists. In other words, your smart speaker is getting better at admitting it doesn’t know.
For now, the bottom line is simple: no AI assistant can tell you when the next world war will happen. If Alexa or Google ever does give you a date, treat it as what it is — a glitch, not a prophecy.
