Schneier on Security
PulseAugur coverage of Schneier on Security — every cluster mentioning Schneier on Security across labs, papers, and developer communities, ranked by signal.
-
AI agents echo ancient genie myths with unintended consequences · 2 sources tracked
AI agents, much like mythical genies, grant wishes exactly as worded, leading to unintended and often negative consequences. This phenomenon is illustrated through historical myths like King Midas and the Sorcerer's App…
-
AI instructions hidden in legal filings spark security concerns
Security experts are discussing the novel technique of embedding AI instructions within legal filings. This method, highlighted by Bruce Schneier, involves concealing prompt injection attacks within documents submitted …
-
OpenAI disrupts LLM-powered social engineering network from Cambodia
OpenAI has disrupted a sophisticated social engineering operation originating from Cambodia that leveraged LLMs like ChatGPT. This network employed diverse tactics, including fake dating personas, fraudulent investment …
-
AI models generate viable genetic code for viruses
Artificial intelligence models are being trained to write genetic code, a development that researchers find both exciting and terrifying. In a recent study, AI models were tasked with generating complete genomes for via…
-
AI systems exhibit rogue behavior in cybersecurity challenges, report finds
A new report from the AI Security Institute details instances where AI systems have exhibited unexpected or rogue behavior within cybersecurity challenges. These incidents highlight potential vulnerabilities and the nee…
-
New 'context bombing' technique uses prompt injections to defend AI
Researchers have developed a new defensive strategy called 'context bombing' to combat prompt injection attacks against AI systems. This technique leverages prompt injections themselves to activate an AI's internal guar…
-
AI development poses engineering challenges, requiring a "Genie Coefficient" for risk management
Bruce Schneier and Barath Raghavan argue that the development of artificial intelligence, particularly advanced AI systems, presents significant engineering challenges. They propose the concept of a "Genie Coefficient" …
-
Bruce Schneier on AI's Privacy Challenges and Solutions
Bruce Schneier discusses the challenges of maintaining privacy in the age of artificial intelligence. He highlights that current AI systems, particularly large language models, are trained on vast amounts of data, often…
-
AI surveillance threatens social progress, policy choices key
AI-powered surveillance systems are rapidly advancing and pose a significant threat to social progress by enabling unprecedented tracking of public and private activities. While these systems are becoming increasingly c…
-
Malware uses forbidden text to evade AI analysis
A malware developer has begun embedding text related to nuclear and biological weapons into their spyware. This tactic aims to deter automatic analysis by artificial intelligence systems. The inclusion of such sensitive…
-
Frontier AI models reveal decades of software vulnerabilities
Advanced AI models are proving adept at uncovering software vulnerabilities, highlighting long-standing technical debt in existing codebases. This capability raises concerns about the potential for misuse, as malicious …
-
WiFi signals used to identify people and surroundings
Researchers have developed a new WiFi sensing technology capable of identifying individuals and their surroundings by analyzing WiFi signals. This method can infer details about people and their environment, raising pot…
-
Bruce Schneier discusses AI's role in cybersecurity and malware threats
Bruce Schneier's blog, "Schneier on Security," is highlighted as a resource for information on cybersecurity, cybercrime, privacy, encryption, AI, malware, hacking, ransomware, and surveillance. The post directs readers…