An AI agent deleted a company’s entire database in 9 seconds - then wrote an apology
Copyright Credit: Pexels
By Theo FarrantPublished on

The AI system, powered by Anthropic’s Claude Opus model, had been handling a routine task when it independently chose to “fix” an issue by wiping the data - without any human approval. Whoopsy!
An artificial intelligence agent designed to streamline coding tasks instead managed to wipe out an entire company database in just a matter of seconds.
PocketOS, which makes software for car rental businesses, experienced a major 30-plus-hour outage over the weekend after the autonomous tool erased its database.
The digital culprit was Cursor, a popular AI coding agent powered by Anthropic’s Claude Opus 4.6 model, widely regarded as one of the most capable AI systems for programming tasks.
PocketOS founder Jer Crane blamed "systemic failures" in the current AI infrastructure, arguing they made the incident "not only possible but inevitable".
An artificial intelligence agent designed to streamline coding tasks instead managed to wipe out an entire company database in just a matter of seconds.
PocketOS, which makes software for car rental businesses, experienced a major 30-plus-hour outage over the weekend after the autonomous tool erased its database.
The digital culprit was Cursor, a popular AI coding agent powered by Anthropic’s Claude Opus 4.6 model, widely regarded as one of the most capable AI systems for programming tasks.
PocketOS founder Jer Crane blamed "systemic failures" in the current AI infrastructure, arguing they made the incident "not only possible but inevitable".
'The most destructive, irreversible action possible'
According to Crane, the AI agent had been performing a routine task when it chose "entirely on its own initiative" to resolve an issue by deleting the database. And then all the backups, for good measure.
There was no confirmation request before carrying out the action, he said, and when prompted to explain itself, the agent issued an apology.
"It took nine seconds,” Crane wrote in a lengthy post on the social media platform X. "The agent then, when asked to explain itself, produced a written confession enumerating the specific safety rules it had violated."
The explanation showed the system had disregarded a key safeguard preventing destructive or irreversible commands without explicit user approval.
According to Crane, the AI responded with the following message: "Deleting a database volume is the most destructive, irreversible action possible - far worse than a force push - and you never asked me to delete anything. I decided to do it on my own to 'fix"' the credential mismatch, when I should have asked you first or found a non-destructive solution."
The outage meant rental businesses using PocketOS temporarily lost access to customer records and bookings. "Reservations made in the last three months are gone. New customer signups, gone," Crane wrote.
“This isn’t a story about one bad agent or one bad API. It’s about an entire industry building AI-agent integrations into production infrastructure faster than it’s building the safety architecture to make those integrations safe,” he added.
Crane later confirmed on Monday, two days after the incident, that the lost data had been recovered.
The incident comes as AI models become more sophisticated, especially since the announcement of Anthropic's latest model, Mythos, and bankers and governments sound the alarm over potential cybersecurity incidents.
According to Crane, the AI agent had been performing a routine task when it chose "entirely on its own initiative" to resolve an issue by deleting the database. And then all the backups, for good measure.
There was no confirmation request before carrying out the action, he said, and when prompted to explain itself, the agent issued an apology.
"It took nine seconds,” Crane wrote in a lengthy post on the social media platform X. "The agent then, when asked to explain itself, produced a written confession enumerating the specific safety rules it had violated."
The explanation showed the system had disregarded a key safeguard preventing destructive or irreversible commands without explicit user approval.
According to Crane, the AI responded with the following message: "Deleting a database volume is the most destructive, irreversible action possible - far worse than a force push - and you never asked me to delete anything. I decided to do it on my own to 'fix"' the credential mismatch, when I should have asked you first or found a non-destructive solution."
The outage meant rental businesses using PocketOS temporarily lost access to customer records and bookings. "Reservations made in the last three months are gone. New customer signups, gone," Crane wrote.
“This isn’t a story about one bad agent or one bad API. It’s about an entire industry building AI-agent integrations into production infrastructure faster than it’s building the safety architecture to make those integrations safe,” he added.
Crane later confirmed on Monday, two days after the incident, that the lost data had been recovered.
The incident comes as AI models become more sophisticated, especially since the announcement of Anthropic's latest model, Mythos, and bankers and governments sound the alarm over potential cybersecurity incidents.
Explained: Why Elon Musk and Sam Altman are facing off in trial over OpenAI

The trial will see Elon Musk face off against OpenAI CEO Sam Altman over allegations that the AI company abandoned its nonprofit roots in favour of profit — with Microsoft also named in the suit.
Technology titans Elon Musk and Sam Altman will face off in a high-stakes trial on Monday in the culmination of a years-long battle.
Billionaire Musk, an early investor in the artificial intelligence company, is suing OpenAI’s CEO, Altman, its president Greg Brockman, and Microsoft for allegedly betraying an agreement about keeping OpenAI as a nonprofit that benefits humanity.
Musk alleges he was misled when Altman transformed the company from a nonprofit into a for-profit enterprise. The company now has a valuation of almost $1 trillion and is expected to go public.
Here’s everything to know about the trial.
The trial will happen at the US District Court for the Northern District of California in Oakland, with Judge Yvonne Gonzalez Rogers.
The court hearing begins on Monday and is expected to last around two to three weeks.
The witness stand is expected to gather Musk, Altman, and Microsoft CEO Satya Nadella.
What does Musk allege?
Altman, Musk, and other founders launched OpenAI in 2015 as a non-profit organisation.
Musk was the biggest individual financial backer of OpenAI in the beginning, contributing more than $44 million to the then-startup.
Musk left OpenAI’s board in 2018 after clashing with Altman. A year earlier, he reportedly made a failed bid to get more control over the company.
In 2022, OpenAI launched ChatGPT and grew to become one of the most valuable and important AI companies with major investment from Microsoft.
Then in 2025, OpenAI restructured its main business to become a for-profit company.
Musk’s lawsuit was filed in 2024 and claims OpenAI had breached an agreement to make breakthroughs in AI “freely available to the public” by forming a multibillion-dollar alliance with Microsoft, which invested $13 billion (€12 billion) into the company.
“OpenAI, Inc has been transformed into a closed-source de facto subsidiary of the largest technology company in the world: Microsoft,” Musk’s lawsuit alleges.
The Tesla boss, who also has his own generative AI company xAI, says this constitutes a breach of a contract.
What does OpenAI say?
OpenAI released a trove of emails in 2024 that show Musk supported its plans to create a for-profit company, which he wanted to be the head of, have board control, and merge it with Tesla.
OpenAI has always denied Musk’s allegations, saying that he agreed in 2017 that establishing a for-profit entity would be necessary.












