駭客智勝 Instagram 人工智慧聊天機器人，劫持帳號！

Instagram recently faced a serious security issue involving its AI chatbot, which allowed hackers to take over other users' accounts. The AI chatbot is designed to help users with account problems, but some hackers discovered a way to trick it. By pretending to be in the same location as the real account owner, hackers could ask the chatbot to change the email address linked to the account. Once the email was changed, the hackers received a verification code at their own email address, which let them reset the password and gain full control of the account. This left the real owners locked out and worried about their personal information.

Meta, the company that owns Instagram, responded quickly to the problem. Andy Stone, a spokesperson for Meta, said that the issue has been fixed and that the company is working to secure all accounts that were affected. He also addressed rumors that world leaders' accounts had been hacked, saying there was no evidence to support those claims. However, one high-profile account used by former US president Barack Obama was reportedly taken over, and hackers posted messages supporting Iran before the account was recovered. It is still unclear how many accounts were actually affected by this exploit.

Jane Manchun Wong, a security researcher and former Meta employee, was one of the people whose account was hacked. She said her password was changed without her knowledge and that she saw several attempts to reset her password. This made her and others concerned about the safety of their accounts. Videos shared on social media showed how hackers used a VPN to pretend to be in the same location as the account owner. They then asked the AI chatbot to link a new email to the account and send a verification code. Once they had the code, they could change the password and take over the account.

Experts warn that using AI chatbots for important tasks like account recovery can be risky if there are not enough security checks. Marijus Briedis, the chief technology officer at NordVPN, said that when chatbots have too much authority and not enough verification, they can become a serious security risk. He explained that account recovery should not be too easy, because the person asking for access might not be the real owner. Many users have complained about the lack of human support when their accounts are hacked or suspended by mistake. In Europe, an independent group that helps social media users said that many people have been banned from their accounts by mistake, even though companies spend billions on AI.

This incident highlights the challenges of using AI in customer service. While AI can make things faster and more convenient, it can also create new risks if not managed carefully. Companies like Meta need to make sure their AI tools are secure and that users can get help from real people when needed. Instagram is working to fix the problem and protect users, but this event shows that technology is not perfect and that human support is still important for keeping people safe online.