On the 31st of last month, a suspicious post appeared on former U.S. President Obama’s official White House Instagram account. The post featured an AI-generated image of a Middle Eastern man alongside the text, “The White House is under the control of the Shia faction.” This account, used by former President Obama during his tenure, had not posted anything since January 2017, when U.S. President Donald Trump assumed office. It was clearly a hacking incident.
TechCrunch and other foreign media reported on the 1st (local time) that the hack occurred due to Meta’s AI customer support chatbot. Hackers requested the AI chatbot to add an email linked to the account, and the AI complied without suspicion, allowing the hackers to infiltrate the system.
According to foreign reports, the hacking method was simple. The hackers used a virtual private network (VPN) to pose as specific individuals and contacted Meta’s AI customer support chatbot. When they asked the AI chatbot to add a new email address to the account, the AI sent a verification code to that email. The hackers shared the code with the chatbot, which then displayed a “password reset” button. This allowed the hackers to input a new password and take over the account—no advanced hacking skills were required.
Using this method, the hackers stole Instagram accounts of celebrities. Among the compromised accounts were former President Barack Obama’s White House account, the account of John Bentivegna, the senior enlisted advisor of the U.S. Space Force, and the account of the cosmetics retailer Sephora. The hackers reportedly sold access to these high-profile accounts for small sums.
Meta expanded AI customer support for Facebook and Instagram in March of this year, allowing the AI to handle core account management functions like password resets. However, the AI failed to perform proper account security checks, and this vulnerability was reportedly shared among hacker groups since March. Users who had their accounts stolen through the AI support system were unable to even file complaints with human agents.
As the issue escalated, Meta took post-incident measures, such as deleting posts made on the hacked accounts. Andy Stone, Meta’s vice president of communications, stated, “The issue has been resolved, and we have addressed the root cause to prevent the AI support system from being exploited in the same way again.”
The tech industry views this incident as a demonstration of how AI with excessive authority can be vulnerable to hacking. The hack succeeded simply through polite requests to the AI, without phishing, malware, or password cracking. A tech industry insider remarked, “This is a case showing that granting too much power to AI can lead to problems.”
Originally written by: Seongmin K
Image credit: Reuters-Yonhap
Source: The Chosun Daily
Published on: 2 June 2026
Link to original article: Obama’s Instagram Hacked via Meta’s AI Chatbot Flaw
