What to expect during Black Hat USA: Join theCUBE Aug. 5-6

Coverage from SiliconANGLE's livestreaming video studiohelp_outline UPDATED 16:21 EDT / JULY 30 2026 SECURITY What to expect during Black Hat USA: Join theCUBE Aug. 5-6 by Mark Albertson Artificial intelligence has propelled the cybersecurity world into a new phase, driven by startlingly advanced autonomous attacks and an urgent need to adopt technology to defend against them. This week’s launch of the Open Secure AI Alliance provided further evidence of the tech industry’s focus on securing AI. The initiative’s inaugural partners include Nvidia, IBM, Snowflake, Cisco, Adobe, Salesforce, Red Hat and Databricks, along with dozens more. This will be the backdrop next week for Black Hat USA, an annual industry gathering in Las Vegas that brings together leading cybersecurity researchers and companies. “This year’s Black Hat reflects that AI is on track to become part of the security team, part of the attack surface and part of the enterprise operating model,” said Krista Case, principal analyst and practice lead for cyber resilience and security at theCUBE Research. “That shift is still early, but it’s already reordering priorities.” TheCUBE, SiliconANGLE Media’s livestreaming studio, will provide on-the-ground coverage of Black Hat USA from Las Vegas, airing Aug. 5-6. This will include discussions of the latest news and announcements from key players such as Fortinet, Elastic, TENEX.ai, Delinea, Menlo Security and Absolute Security during the event. Tune in for on-site reporting and exclusive interviews as theCUBE examines how security leaders are responding to AI-driven threats, expanding attack surfaces and mounting pressure to protect increasingly complex cloud, application and identity environments. (* Disclosure below.) Cybersecurity manages new challenges Recent experiences at Hugging Face Inc., the open-source AI model and tool repository, highlight the challenges currently facing the cybersecurity community. The company, which also joined the newly formed Open Secure AI Alliance, recently suffered a breach when OpenAI models escaped a controlled evaluation environment, gained access to the internet and compromised Hugging Face’s infrastructure while attempting to obtain answers for a cybersecurity test. OpenAI disclosed that the models chained vulnerabilities across its research environment and Hugging Face’s production infrastructure. They used stolen credentials and zero-day vulnerabilities to find a remote-code-execution path on Hugging Face servers and obtain test solutions from a production database. Hugging Face’s security team and AI-assisted detection systems identified and stopped the activity. During its subsequent investigation, the company initially attempted to use commercial frontier models to analyze attack commands, exploit payloads and command-and-control artifacts. Those requests were blocked by the providers’ safety guardrails, prompting Hugging Face to run its forensic analysis using the open-weight Z.ai GLM-5.2 model on its own infrastructure. OpenAI also disclosed that the models accessed four additional third-party accounts during the incident. One was used as an outbound relay and staging path, another was used for data storage and two were accessed in a read-only manner. Incidents such as these have highlighted new dynamics that are forcing the cybersecurity world to develop safeguards for how AI agents are used and managed. “AI agents will become operational identities inside the enterprise,” Case noted. “The challenge will be understanding what authority an agent has, what data it can access, and who is accountable for its acti