According to security researchers at Metomic, 40.2% of files stored on Google Drive contain sensitive data.  The researchers analyzed roughly 6.5 million Google Drive files.  The researchers also found that 34.2% of the scrutinized files were shared with external contacts outside the company’s domain.  Equally concerning, the researchers noted that over 350,000 files (0.5%) were publicly accessible, allowing unrestricted entry to anyone possessing the document link.

Threat actors could use a critical Bluetooth security flaw to take control of Android, Linux, macOS, and iOS devices. The vulnerability, tracked as CVE-2023-45866, is a case of authentication bypass that allows attackers to connect to vulnerable devices and inject keystrokes to conduct code execution as the victim. According to security researcher Marc Newlin, who disclosed the issue to software vendors in August 2023, multiple Bluetooth stacks have authentication bypass flaws that enable attackers to connect to a discoverable host without user confirmation as well as inject keystrokes.

Researchers have shown that most popular password managers can leak credentials from Android devices when using the mobile operating system's WebView autofill function with malicious apps. Ankit Gangwal of the International Institute of Information Technology (IIIT) demonstrated how mobile apps that use WebView controls can leak passwords from many password managers. In a paper, Gangwal and his students Shubham Singh and Abhijeet Srivastava detail the "AutoSpill" credential-leaking vulnerability.

Artificial Intelligence (AI) security researchers from Robust Intelligence and Yale University have developed a Machine Learning (ML) method that can quickly jailbreak Large Language Models (LLMs) and do so in an automated way. According to Robust Intelligence researchers, the Tree of Attacks with Pruning (TAP) method can be used to induce sophisticated models such as GPT-4 and Llama-2 to generate hundreds of harmful responses to a user query in minutes. Their findings imply that this vulnerability is widespread in LLM technology, but they do not see an obvious solution.

Cybersecurity data breaches present unique risks for victim-survivors of domestic violence. For example, the UK Information Commissioner reported earlier this year that it had reprimanded seven organizations for privacy breaches involving victims of domestic abuse. These included organizations revealing the victims' safe addresses to their alleged abusers. In one case, a family had to be relocated to emergency housing immediately. In another case, an organization revealed two children's home address to their abusive birth father.

According to security researchers at SecurityScorecard, almost all (90%) of the world’s 48 biggest energy companies have suffered a supply chain data breach in the past 12 months.  The researchers analyzed the cybersecurity posture of the largest coal, oil, natural gas, and electricity companies in the US, UK, France, Germany, and Italy, as well as their suppliers,  covering 21,000 domains.  The researchers identified 264 breach incidents related to third-party compromises in the past 90 days alone.  The researchers noted that some countries fared better than others.

Researchers from the Systems and Network Security Group (VUSec Group) at Vrije Universiteit Amsterdam have developed "Spectre based on LAM" (SLAM), a new side-channel attack that uses hardware features designed to improve security in upcoming Intel, AMD, and Arm CPUs to obtain the root password hash from kernel memory. According to the researchers, SLAM is a transient execution attack that exploits a memory feature, which enables software to use untranslated address bits in 64-bit linear addresses for storing metadata.

As part of a new research project, Dr. Rosetta Romano, an assistant professor in Information Technology (IT) and Systems at the University of Canberra, will help small businesses by educating them about the criticality of cybersecurity. Dr. Romano and fellow researchers, together with the Canberra-based data management company Surround Australia, are developing a Cybersecurity Standards and Frameworks Knowledge Graph for Australian small businesses. Large companies tend to have greater resilience and ability to continue operations in the event of a cyberattack.

Synthetic data can be used to improve cybersecurity algorithms, such as facial recognition and fingerprint matching. These applications have the potential to improve suspect identification in law enforcement and streamline border control processes. Synthetic data generation involves manipulating variables and parameters in order to replicate data patterns. Since synthetic data does not contain real individuals' sensitive information, it can help to alleviate data privacy concerns. This is important in criminal investigations, where privacy is critical.

The National Security Agency (NSA) has joined the Cybersecurity and Infrastructure Security Agency (CISA) and other US and international partners in releasing a Cybersecurity Information Sheet (CSI) titled "The Case for Memory Safe Roadmaps." The report, which expands on NSA's "Software Memory Safety" CSI published in April 2023, offers guidance for software manufacturers and technology providers in developing roadmaps tailored to eliminate memory safety vulnerabilities from their products.