SANS AI security summit

SANS Institute is staging an artificial intelligence cybersecurity summit on April 20 and 21 as security teams race to adapt to machine-made attacks and defenses. (sans.org) The event page lists a summit on April 20 and 21, followed by training through April 27, with in-person sessions at Hilton Arlington Rosslyn and a virtual option in Eastern Time. SANS says the event includes nine courses and offers 12 continuing professional education credits for summit attendees. (sans.org) SANS’ broader summit program says its cybersecurity summits are free to attend and globally accessible, while the artificial intelligence event also includes a separate two-day “Solutions Track” on April 20 and 21. That track is chaired by Matt Bromiley and focuses on threat detection, response automation, and analyst decision-making with human oversight. (sans.org, sans.org) Artificial intelligence in cybersecurity usually means two different things at once: defenders use models to sort alerts and automate repetitive work, while attackers use the same kind of systems to find weaknesses and speed up intrusions. SANS says its summit sessions will cover both deploying artificial intelligence inside security operations and defending against artificial intelligence-powered threats. (sans.org, sans.org) That framing follows a sharper tone from SANS and its partners this week. In a joint April 14 briefing announcement, SANS and the Cloud Security Alliance said artificial intelligence-driven vulnerability discovery is compressing exploit timelines from “weeks to hours,” and said more than 60 contributors and 250 chief information security officers helped assemble a response guide. (sans.org, cloudsecurityalliance.org) The same briefing lays out why security leaders are moving quickly. It cites XBOW reaching HackerOne’s United States leaderboard in June 2025, the Defense Advanced Research Projects Agency’s Artificial Intelligence Cyber Challenge finding 54 vulnerabilities in 54 million lines of code in four hours in August 2025, and Anthropic reporting more than 500 high-severity open-source vulnerabilities in February 2026. (cloudsecurityalliance.org) The summit agenda reflects a second pressure point inside companies: employees are already using generative artificial intelligence tools in everyday software. One April 20 session description says security teams now have to govern “sanctioned and shadow” artificial intelligence in tools such as Slack, Microsoft 365, Claude, and Grammarly, rather than trying to block everything outright. (sans.org) SANS is also carving out closed-door discussion for senior executives. An invitation-only session on April 21 is set aside for chief information security officers and other leaders to discuss governance, ethics, and security in a moderated format with no speeches or panels. (sans.org) The immediate test comes next week, when SANS tries to turn a fast-moving threat picture into operating advice that analysts, engineers, and executives can use on April 20 and 21. (sans.org, sans.org)