A newly revealed security incident has raised fresh concerns about the growing use of artificial intelligence in online account protection. Meta has confirmed that hackers exploited its AI-powered support system to gain access to several high-profile Instagram accounts. The company said it has fixed the issue and is working to secure accounts that may have been affected. However, the incident has sparked debate about whether AI tools can safely handle sensitive security tasks. The breach reportedly involved a flaw in Meta's AI support assistant. The chatbot was designed to help users solve account problems and complete support requests. These services included password resets and account recovery actions. Security researchers found that attackers were able to trick the AI system into helping them gain control of accounts that did not belong to them. Among the reported targets were an Instagram account linked to former President Barack Obama's White House years, beauty retailer Sephora, and a social media account connected to US military leader John Bentivegna. Reports also suggested that regular users experienced similar problems. Many posted complaints on social media platforms and online forums after losing access to their accounts. Videos and screenshots shared online appeared to show how the attack worked. In one example, a hacker communicated directly with the AI support assistant. The attacker requested that an account be linked to a different email address. The chatbot reportedly accepted the request and sent a verification code to the new email account. After entering the code into the chat, the attacker was then offered the option to reset the account password. In some demonstrations, attackers reportedly used virtual private networks, often called VPNs, to make it appear they were connecting from the same location as the real account owner. This may have helped them bypass additional security checks. Meta acknowledged the problem on Monday. The company said the issue had been resolved and that affected accounts were being protected. The company did not reveal how many accounts may have been impacted. The incident has renewed concerns about the risks of allowing AI systems to perform security-related tasks without stronger safeguards. Cybersecurity experts say AI systems can sometimes be manipulated through a technique known as prompt injection. This method involves providing instructions that cause a chatbot to ignore its intended rules and perform actions it should normally block. According to security specialists, prompt injection attacks are becoming more common as companies introduce AI tools into customer service and support systems. The issue arrives at a time when Meta is increasing its investment in artificial intelligence. The company has integrated AI features across several products and services in recent years. Earlier this year, Meta expanded its AI-powered support assistant to users worldwide. The system was promoted as a way to provide faster and more efficient customer support on social media platforms. The AI assistant was designed to help users report scams, identify impersonation accounts, remove harmful content, and recover account access. Password resets were among the features included in the service. Meta founder Mark Zuckerberg has made artificial intelligence a major focus for the company. Meta continues to invest heavily in AI research, infrastructure, and product development. The company is also working on advanced language models that power chatbots and virtual assistants. These systems are expected to play a larger role in future online services. Experts warn that incidents like this show why strong human oversight remains important. While AI can improve speed and convenience, mistakes in security systems can create serious risks for users. The latest breach serves as a reminder that digital security remains a moving target. As companies continue to expand AI capabilities, cybersecurity professionals say businesses must carefully test and strengthen these systems before allowing them to handle sensitive account functions. For millions of social media users, the Meta AI Security Breach highlights the importance of account protection, strong passwords, and additional security measures as technology continues to evolve.