Overview
SecureAI.se is an AI security event held at Nalen in Stockholm from 10 to 12 November 2026. It combines two days of small-group, hands-on training on 10 and 11 November with a single-track conference day on 12 November. The organisers describe it as the second annual deep-dive into AI security, covering both offensive and defensive topics. Last year's attendees came from Fortune 500 companies, banks and financial institutions, and government organisations.
Who should attend
The event is aimed at security professionals, engineers, and decision makers who work with AI systems in production, including those in regulated sectors such as banking, finance, and government.
Format
The trainings are instructor-led and limited to small groups, so participants build, break, and defend real AI systems rather than only watching slides. Every training ticket includes the full conference day on 12 November. The conference day runs as a single track with keynotes and talks, plus networking events and lunch.
Training topics
- Training LLMs for Security Engineers (10 to 11 November, instructor Thomas Olofsson): a two-day hands-on course on building and fine-tuning security models with real GPUs and LoRA/QLoRA. Participants create log anomaly detectors and vulnerability scanners, then choose tool automation or detection rule generation for a capstone project, using models designed to run in their own environment.
- Pentesting LLM Applications and Agents (10 to 11 November, instructor Donato Capitella of Reversec): a two-day course on practical LLM pentesting using Spikee, a mailbox summariser, and an agentic banking chatbot. Participants learn to automate prompt-injection attacks, test memory and tool authorization, evaluate guardrails, and report reproducible evidence.
Agenda (12 November 2026)
- 09:00 Keynote: High volume AI vulnerability reporting, Daniel Stenberg (curl)
- 10:05 Model x Harness x Alignment, Shaun Smith (Hugging Face)
- 10:45 Playing CTF in prod with agents, Movitz Sunar (Lovable Labs)
- 11:45 Talk: to be announced
- 13:00 Verifiable Reasoning Architecture for Security Agents, Ron F. Del Rosario (SAP)
- 13:00 Building AI Software Factories and Keeping Them Secure, Lezgin Zilan (Billogram), in Harlem
- 13:40 Does Your AI Leak? Measuring Privacy Risk in Production Models, Fazeleh Hoseini (AI Sweden)
- 14:20 The State of Supply Chain SlopResponse, Rami McCarthy (Wiz, Google)
- 15:20 From AI Regulation to Secure AI Systems, Luis Martinez (TELETID)
- 16:00 Kryptonite for Agents: Trapping AI Agent Swarms in a Honeypot, Krister Hedfors (Basalt)
- 16:25 National mitigation of AI driven cyberthreats, Tommy Schönberg (AI Sweden) and Camilla Lundahl (Cyrenity)
- 16:50 Closing remarks
The programme includes scheduled coffee breaks and lunch.
Speakers
The lineup includes Daniel Stenberg, founder and lead developer of the curl project; Krister Hedfors, Director of AI at Basalt; Lezgin Zilan, CTO at Billogram; Rami McCarthy, Principal Security Researcher at Wiz (Google); and Ron F. Del Rosario, VP and Head of AI Security at SAP Supply Chain Management. Other speakers include Fazeleh Hoseini of AI Sweden, Luis Martinez, Shaun Smith of Hugging Face, and Movitz Sunar of Lovable Labs. Thomas Olofsson and Mattias Bååth moderate the event.
Topics
The wider programme covers threat modeling for AI-driven systems, responsible AI development, data privacy and compliance (including GDPR and data residency), secure MLOps pipelines, AI regulation such as the EU AI Act, and infrastructure security for on-premise and cloud AI compute.
Tickets
- Conference ticket: 6,000 SEK excluding VAT (25% VAT added), with a discounted government and public sector rate of 5,000 SEK excluding VAT that requires valid ID.
- Training with conference ticket included: 21,000 SEK excluding VAT (25% VAT added).
- Group discounts are offered: 15% off for 3 or more tickets and 20% off for 5 or more.
- Early bird pricing is advertised, and training seats are limited.




