Cloud security provider Wiz found two critical architecture flaws in generative AI models uploaded to Hugging Face, the leading hub for sharing AI models and applications. In a blog post published on April 4, Wiz Research described the two flaws and the risk they could pose to AI-as-a-service providers. These are: Shared Inference infrastructure takeover […]
98% of companies surveyed view some of their AI models as vital for business success, and 77% have experienced breaches in their AI systems over the past year, according to HiddenLayer. The report surveyed 150 IT security and data science leaders to shed light on the biggest vulnerabilities impacting AI today, their implications for commercial […]
98% of companies surveyed view some of their AI models as vital for business success, and 77% have experienced breaches in their AI systems over the past year, according to HiddenLayer. The report surveyed 150 IT security and data science leaders to shed light on the biggest vulnerabilities impacting AI today, their implications for commercial […]
Generative AI based on large language models (LLMs) has become a valuable tool for individuals and businesses, but also cybercriminals. Its ability to process large amounts of data and quickly generate results has contributed to its widespread adoption. AI in the hands of cybercriminals According to a report from Abnormal Security, generative AI (GenAI) is […]
As large language models (LLMs) become more prevalent, a comprehensive understanding of the LLM threat landscape remains elusive. But this uncertainty doesn’t mean progress should grind to a halt: Exploring AI is essential to staying competitive, meaning CISOs are under intense pressure to understand and address emerging AI threats. While the AI threat landscape changes […]
LLM Guard is a toolkit designed to fortify the security of Large Language Models (LLMs). It is designed for easy integration and deployment in production environments. It provides extensive evaluators for both inputs and outputs of LLMs, offering sanitization, detection of harmful language and data leakage, and prevention against prompt injection and jailbreak attacks. LLM […]
The new MK3 models feature polyphonic aftertouch. [embedded content] Native Instruments have announced updates to its Komplete Kontrol S range of MIDI keyboards. The MK3 model features a number of major upgrades and a new design, but the biggest news is the introduction of polyphonic aftertouch, a rare and highly sought after feature found on […]
The true stereo Eurorack module features three delay models and 70s-inspired reverb. [embedded content] Intellijel’s latest announcement is the Sealegs delay module, a digital design intended to be “the warmest, most musical and organic delay module possible”. The module features three delay models (tape, BBD and crossfading digital), full stereo operation, up to eight seconds […]
Four Iteritas models are set for the chop. Fare thee well, sweet Basimilus. [embedded content] Noise Engineering have announced that the four core modules in the Iteritas series will be discontinued. The Basimilus Iteritas Alter, Manis Iteritas, Cursus Iteritas and Ataraxic Iteritas all rely on an XMOS CPU chip which has been taken out of […]