Students endure a simulated earthquake during the annual Great California ShakeOut earthquake drill on Oct. 16, 2014.
On Monday, a bot led to incorrect information going out on Twitter. That’s hardly unprecedented. But this case was different: It was the Los Angeles telling the public there had been a 6.8 magnitude earthquake in Isla Vista, California, even though no such quake had taken place.
The tweet was prompted by an alert from the U.S. Geological Survey that said the earthquake happened on June 29, 2025. But it was actually referring to a real earthquake that happened almost a century ago.
So how did this happen?
A system at USGS records ground motion, analyzes is compared to other motion in the area to detect where it’s coming from, and then declares if an earthquake is occurring. This week, researchers tried to update the location of an earthquake from 1925. Thanks to a bug with the software that sends out email updates, subscribers received an email notification telling them an earthquake had occurred in 2025.
The Los Angeles Times Quakebot received this information and did what it’s designed to do: It wrote up basic information about the quake (location and magnitude) and then tweeted it out. As Will Oremus explained in Slate in 2014, the bot is in place not to replace real, live journalists, but to make it easier to release quick information about emergency situations.
However, this small mistake led some to think an actual, fairly large earthquake had taken place. Had that had been true, that’s guaranteed to change your day around a bit, especially if you’re a local reporter.
Soon enough people were taking to Twitter to ask what was going on. The Los Angeles Times quickly deleted the tweet and published an article explaining what had occurred.
So does this flub suggest that we should stop depending on bots to alert people to earthquakes? Lucy Jones, a seismologist and USGS scientist emerita, doesn’t think so.
Previously, updates would have to be run by a person, and Jones said that takes too much time in the case of a real emergency.
“The only way to not have misinformation is to not have information. [The bot] is completely accurate if the USGS email’s accurate,” she said. “This is the first mistake of this type that we’ve seen … and now it’s been fixed.” She thinks the bot is still worthwhile, particularly because when we expect flawless information, we stop thinking about where that data came from.
“I’m a little concerned about how people think information just sort of magically appears,” Jones said. “We get so accustomed to the weather showing up on our watch now, you don’t think about the billions of dollars of satellites and people and computers that go in to do weather modeling.”
Thankfully, this mistake doesn’t seem like a career ending move for the Los Angeles Times bot. Hopefully other early detection systems can also stay in place, like the tsunami monitoring stations in the Pacific Ocean, which this administration seeks to cut U.S. funding for.
Future Tense is a partnership of Slate, New America, and Arizona State University.