Generative-AI security has moved to the forefront of this year’s Black Hat USA conference, where the Open Worldwide Application Security Project’s newly established GenAI Security Project is setting guidelines for developers and corporate security teams. Vendors are rushing in with dedicated products. SentinelOne said it will acquire Tel-Aviv–based Prompt Security to fold the start-up’s guardrails for large language models into its Singularity platform. Tenable introduced “AI Exposure,” a risk-assessment module for enterprises deploying generative models, while Sysdig launched what it calls the industry’s first agentic cloud-security platform. Toronto-based Cohere took the wraps off its North productivity suite, making the agent-based platform generally available after a seven-month pilot with Dell Technologies and the Royal Bank of Canada. North can run behind customers’ firewalls on as few as two GPUs and incorporates access controls, continuous red-teaming and compliance with GDPR, SOC-2 and ISO 27001, the company said. Additional releases include Absolute Security’s near-real-time GenAI assistant for endpoint protection and Black Duck’s upgraded AI-powered code-analysis assistant, signalling that both traditional and emerging security vendors see a widening addressable market. The product blitz follows research from Palo Alto Networks and CrowdStrike showing that adversaries now harness generative AI to automate phishing and intrusion workflows, compressing attack lifecycles from days to minutes and driving demand for purpose-built defences.