Anthropic Threat Intelligence Report - Sept 2026 : DECODED

Speaker - Anshu Gupta Anthropic's September 2026 threat intelligence report runs over 100 pages and covers seven distinct harm categories, from AI-orchestrated Russian espionage to fake election-influence platforms. Most security leaders won't get past the executive summary before the next fire drill pulls them away, and that's exactly the problem this session solves. In this session, we'll decode the report's most important findings into what actually matters for your defense posture. You'll leave with a clear picture of how threat actors are now using AI as an autonomous orchestrator rather than a simple assistant, closing the skills gap that used to separate lone hackers from nation-state teams. We'll walk through real case studies, including a Russian espionage group that used AI to detect and automatically rebuild its own malware when flagged by security tools, and a wave of attacks where stolen AI API keys became the new high-value credential, giving attackers compute, cover, and loot in a single theft. This isn't a summary. It's a working session designed to translate 8 months of disrupted threat activity into a prioritized action list: what to monitor, what to ask your vendors, what to change in your incident response playbooks, and where your current detections are likely blind to AI-augmented tradecraft. Who should attend: CISOs, security architects, threat intel analysts, and incident responders who need the signal without wading through the full report themselves. What you'll walk away with: