Tags → #generative ai

Tools
Innovative Watermarking Scheme Developed for Generative Image Models
Nov 20, 2024 A new undetectable watermarking scheme for generative image models has been developed, ensuring robustness and image quality while addressing challenges in AI-generated content.
Threats
Jailbreaking Techniques and Security Risks in Large Language Models
Nov 17, 2024 The practice of jailbreaking Large Language Models (LLMs) involves manipulating AI systems to produce unintended outputs, raising concerns about potential security vulnerabilities and prompting research into enhancing their security.
Tools
Study Highlights Advances in Pseudorandom Codes for AI Security
Nov 12, 2024 Pseudorandom codes (PRCs) are advanced error-correcting codes that enhance data integrity and security, particularly in watermarking AI-generated content and protecting against adversarial attacks.
Tools
Researchers Introduce Pseudorandom Codes for Enhanced Cybersecurity
Nov 11, 2024 Researchers Christ and Gunn have introduced Pseudorandom Codes (PRCs) in 2024, enhancing error correction and security in cryptography and cybersecurity.
Tools
Generative AI Framework Aims to Streamline Security Control Development
Nov 8, 2024 Recent advancements in Generative AI may significantly reduce the time required to develop security controls for cloud-based services, potentially streamlining the process to under one minute while maintaining quality and reliability.
Tools
WhiteRabbitNeo: High-Powered Potential of Uncensored AI Pentesting for Attackers and Defenders
Nov 2, 2024 WhiteRabbitNeo is an offensive security gen-AI model, he told SecurityWeek. Its purpose is to allow security teams to examine their infrastructures, detect vulnerabilities that can be exploited (by developing exploits for the vulnerabilities), and provide remediations for those vulnerabilities, explains Manoske.