Breaking news

Instagram Secures Platform After AI Chatbot Exploit Leads To Account Hijackings

Instagram Patches Security Vulnerability

Instagram has fixed a security flaw that allowed hackers to take control of user accounts by exploiting Meta’s AI-powered support chatbot. The vulnerability reportedly enabled attackers to add unauthorized email addresses to accounts and reset passwords without access to the legitimate account owner’s email.

Exploit Mechanics Detailed

Reports of account hijackings surfaced over the weekend through posts on Reddit and warnings shared on X. Among the accounts reportedly affected were the Obama-era White House account and the account of U.S. Space Force Chief Master Sergeant John Bentivegna, raising concerns about the potential scope of the vulnerability.

How The Attack Unfolded

Security researcher Jane Wong said her account was compromised after her password was changed without her knowledge. In a post on X, Wong described receiving repeated password reset notifications before losing access to her account. A widely shared demonstration of the exploit showed how an attacker could use a VPN, interact with Meta’s AI Support Assistant and submit an alternative email address. After receiving a verification code, the attacker could reset the password and gain control of the account without accessing the owner’s original email.

Industry Reactions And The Path Forward

Instagram spokesperson Andy Stone confirmed that the vulnerability has been fixed, although Meta has not disclosed how many accounts may have been affected. The incident highlights the security challenges that can emerge as technology companies expand the use of AI-powered support tools and automated account management systems.

Ongoing Security Challenges

The breach has renewed scrutiny of how AI-driven customer support systems handle account recovery and identity verification requests. While the flaw was addressed quickly, the incident demonstrates how automated support processes can become targets for abuse when security controls fail to account for unexpected forms of manipulation.

A New Twitter-Inspired Social Network Is Taking Shape

A new social network called Twitter.now is entering the market, with a founding team that includes former Twitter trademark counsel Stephen Coates. The service is being developed by startup Operation Bluebird.

As Ars Technica reported, X sued the company last year and asked a Delaware judge to block the launch. Operation Bluebird argued in a petition that X had abandoned trademarks including “Twitter” and “Tweet.”

Coates has said the project is not an attempt to recreate the original Twitter. In a LinkedIn post, he described the platform as a new public space focused on trust, transparency and user choice.

AI System To Rate Posts

Twitter.now is currently being tested, with early access priced at $20. Its main feature is VERA, an AI system designed to evaluate posts, verify claims and provide sources and context.

Posts receive a trust score, with users eventually able to set a minimum score to filter their feeds. The company says this approach will give people more control over what they see instead of leaving those decisions entirely to an algorithm.

Moderation Remains A Challenge

Scaling moderation will be one of the platform’s biggest tests. Social networks have repeatedly struggled with content moderation as their communities grow, and newer platforms such as Bluesky have faced similar criticism.

Operation Bluebird says VERA will form the basis of its moderation and verification system. A second version is already planned, with expanded tools that would let users set a specific trust threshold for the posts appearing in their feeds.

For now, Twitter.now remains in an early testing phase, combining the familiarity of the Twitter name with an AI-driven approach to evaluating online information.

The Future Forbes Realty Global Properties
eCredo
Aretilaw firm
Uol

Become a Speaker

Become a Speaker

Become a Partner

Subscribe for our weekly newsletter