A Spanish software engineer accidentally took control of around 7,000 robot vacuums worldwide due to a shocking security vulnerability, which allowed him to access live camera feeds from inside homes.

Sammy Azdoufal, while trying to reconfigure his DJI Romo vacuum to be compatible with his PlayStation 5 controller using artificial intelligence, inadvertently unlocked widespread access to the device’s servers in 24 different countries, he shared with The Verge.

“I discovered my device was just one among a vast number of others,” Azdoufal explained to the New York-based tech publication.

“I didn’t break any rules, nor did I bypass, crack, or use brute force techniques,” he clarified.

Without needing to hack into DJI’s servers, Azdoufal managed to commandeer thousands of these robots, view through their cameras, listen via their microphones, map interior layouts, and track devices using IP addresses, he admitted.

Moreover, the engineer could directly access the tech company’s network—highlighting how easily malicious individuals could potentially exploit data in our tech-heavy world.

Azdoufal, head of AI at a property management and travel group in Spain, said he alerted The Verge to the company’s major security flaw, prompting the outlet to contact DJI.

The Chinese firm, known for its AI-powered gadgets, confirmed the glitch has since been fixed, adding that other network issues will be addressed in the coming weeks.

“DJI can confirm the issue was resolved last week and remediation was already underway prior to public disclosure,” company spokesperson Daisy Kong wrote in a statement to The Verge.

“DJI maintains strong standards for data privacy and security and has established processes for identifying and addressing potential vulnerabilities.DJI will continue to implement additional security enhancement as part of its ongoing efforts.”

The jarring incident comes two years after Chinese-made robot vacuums by Ecovacs were hijacked and rewired to bombard US homeowners with racial slurs.