Researchers have found that artificial intelligence (AI) models can easily decode encrypted data, raising concerns about potential personal data breaches. A team of researchers from the University of California, Berkeley, and the Massachusetts Institute of Technology (MIT) recently published a study on this issue. The researchers tested various AI models, including those using machine learning and deep learning techniques, to see if they could decode encrypted data. They found that some AI models were able to decode encrypted data with surprising accuracy, even when the data was encrypted using secure protocols. The researchers used a type of encryption called "homomorphic encryption," which allows computations to be performed on encrypted data without decrypting it first. However, they found that some AI models were able to bypass this encryption and decode the data with a high degree of accuracy. This raises concerns about the potential for personal data breaches, as encrypted data is often used to protect sensitive information such as financial data and personal identifiable information. The researchers warned that this vulnerability could be particularly problematic for organizations that rely heavily on AI and machine learning. "As AI models become more widespread, it's essential that we develop more robust security measures to protect sensitive data," said one of the researchers. The study's findings highlight the need for further research into the intersection of AI and cybersecurity. The researchers plan to continue exploring this issue and developing new methods for protecting encrypted data from AI-powered attacks. In the meantime, organizations are advised to exercise caution when using AI models to handle sensitive data. They should also consider implementing additional security measures, such as multi-factor authentication and secure data storage, to protect against potential data breaches.
Researchers at ELLIS and Max Planck Institute have found that high-performance AI models from companies like OpenAI, Anthropic, and Google may leak users' personal data. The team discovered a structural vulnerability in encrypted inference blocks, which can be exploited to access sensitive information. They extracted 315,320 encrypted blocks from 6,708 AI agent execution records and recovered 367 personal identification information and 182 authentication credentials.
The researchers have informed the relevant companies, and the vulnerability is no longer exploitable. They suggest improvements, such as storing encrypted blocks on company servers or not reusing them outside of the conversation session.
Written by urgent.news from Hankyoreh's report — not a translation of it. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- AI สร้างตัวตนปลอมหลอกนักพัฒนา — สิ่งที่โปรแกรมเมอร์ต้องรู้และระวัง dev.to
- How much electricity does it cost to ask AI one question? hani.co.kr
- A photo of a disaster victim posing with a peace sign in front of their completely destroyed home has been widely shared. However, it appears to be a fake. The photo, which has been widely shared on social media, shows a woman smiling and holding up a peace sign in front of her home, which appears to have been completely destroyed in a disaster. However, upon closer inspection, it appears that the photo may have been manipulated using artificial intelligence (AI). The photo was first posted on a social media account on the 13th, and it quickly went viral, with many people expressing admiration for the woman's resilience in the face of disaster. However, some people noticed that the photo seemed too perfect, and began to suspect that it was fake. When the photo was analyzed, it was found that the faces and clothes of the people in the background appeared to be generated by AI. Furthermore, the woman in the photo and her family have not been found in the area where the disaster occurred, and it appears that the photo was created by manipulating images of a real destroyed home. The source of the photo is currently unknown, and it is not clear why someone would create and share a fake photo of a disaster victim. The incident highlights the potential for AI-manipulated photos to be used to deceive people, and the need for greater awareness and caution when sharing photos online. asahi.com
