Tags → #artificial intelligence

Tools
AI Integration in DevSecOps Enhances Software Security Management
Nov 22, 2024 The integration of artificial intelligence into DevSecOps is reshaping security management in software development by enhancing threat detection and automating compliance processes.
Vulnerabilities
Study Reveals Vulnerabilities of LLMs to Bit-Flip Attacks
Nov 22, 2024 A recent study reveals that large language models (LLMs) are vulnerable to bit-flip attacks, which can drastically reduce their performance, prompting the introduction of a new framework called AttentionBreaker to identify and exploit these vulnerabilities.
Education
Study Examines Information Security Awareness of Large Language Models
Nov 21, 2024 A study from Ben-Gurion University of the Negev examines the information security awareness of large language models, revealing gaps in their ability to provide safe security advice and highlighting the need for caution in their use.
Vulnerabilities
Study Proposes Framework for Managing AI Vulnerabilities
Nov 20, 2024 A recent paper from the University of Hawaii at Manoa discusses the vulnerabilities in artificial intelligence systems and proposes the creation of an Artificial Intelligence Vulnerability Database to address these issues.
Tools
Text-to-Image Model Verification Method Proposed by Researchers
Nov 20, 2024 Researchers have developed a new verification method called Text-to-Image Model Verification via Non-Transferable Adversarial Attacks (TVN) to address concerns about the authenticity and reliability of third-party Text-to-Image models.
Vulnerabilities
Survey Highlights Security and Privacy Risks of Large Language Models
Nov 19, 2024 A recent survey highlights significant security and privacy threats associated with large language models (LLMs), identifying vulnerabilities such as jailbreaking and data poisoning across various sectors.
Policy
Meta's Llama Model Used in Military Applications and Policy Shift
Nov 17, 2024 Meta Platforms Inc.'s Llama large language model has garnered attention for its use in both commercial and military applications, prompting discussions about its accessibility and implications for national defense.
Industry
Survey Examines AI Adoption in Cybersecurity Industry
Nov 17, 2024 HackerOne and the SANS Institute conducted a survey revealing that 66% of cybersecurity professionals are using AI technologies, with many acknowledging its benefits in threat detection and incident response, while also expressing concerns about ethical implications and potential risks.
Policy
Challenges of Shadow IT in Modern Enterprises
Nov 14, 2024 Shadow IT has become a significant challenge for enterprise IT leaders, particularly with the rise of cloud computing and AI technologies, complicating visibility and control over unsanctioned applications used by employees.
Vulnerabilities
O-RAN's Role in Wireless Communication and Security Challenges
Nov 14, 2024 The Open Radio Access Network (O-RAN) architecture enhances wireless communication systems through increased flexibility and security, while also introducing new vulnerabilities and challenges.