Secure & Sovereign AI Workshop

Secure & Sovereign AI Workshop

If AI is to serve humanity broadly, not just a handful of centralized players, its capabilities need to grow in directions that strengthen security, protect privacy, and allow peaceful cooperation to emerge among a rising plurality of AI systems and humans.

Today, AI development is outpacing our defenses. Traditional security frameworks can’t keep up with the speed, scale, and complexity of AI-driven threats. Privacy-preserving technologies remain fragmented and hard to scale. And as AI systems proliferate into an ecology of coordinating agents, we lack robust mechanisms that favor cooperation over zero-sum competition.

This workshop convenes leading researchers, engineers, and funders to explore how to make AI an engine of defense in a multipolar human-AI world.

Focus Areas

Participants will collaborate across three interconnected domains:

1. AI for Secure AI

To match the speed, scale, and complexity of AI-supported attacks, we need self-improving defense systems where AI autonomously identifies vulnerabilities, generates formal proofs, red-teams, and strengthens our digital infrastructure. 

2. AI for Private AI

If we want AI progress to occur openly without sacrificing privacy, we need to build confidential compute environments, scale encrypted mechanisms for handling data, and design infrastructure that distributes trust. If you’re building a local compute stack, we especially invite you to learn and build together.

3. AI for Decentralized & Cooperative AI

To approach a resilient multipolar, multiagent world, we need to build decentralized intelligence ecosystems where AI systems cooperate, negotiate, and align. Which tools in economics, game theory, mechanism design, and auxiliary fields are uniquely positioned to build this world? 

Participants

Participants include a tightly curated set of ~80 Foresight grantees, fellows and other selected researchers, engineers, cryptographers, security practitioners, and funders working at the intersection of AI and defensive technologies. 

Format

The workshop features brief talks by selected speakers and an unconference-style working group session for project collaborations. Projects incubated at the workshop are eligible for (further) funding by Foresight’s grants, and may be considered by other funders present at the meeting. In addition to short presentations and working groups, we offer mentorship hours, open breakouts, and speaker & sponsor gatherings.

This workshop is part of Foresight Institute’s AI Node program in Berlin, offering funding, compute and community for AI-powered science and safety projects. After the workshop, participants have the opportunity to stay for a few weeks to sprint at the Berlin office and continue building in community.

Why Attend

Your participation should offer:

• New insights supporting your immediate work in AI safety, security, or privacy
• Lasting collaborations with leaders working on shared defensive goals
• Incubation of novel, fundable projects to drive progress at the frontier of AI-first safety

Chatham House Rule

Workshop discussions are held under Chatham House Rule (don’t connect people to ideas when discussing them outside of this workshop), apart from some presentations which will be recorded.

Apply

Applications are reviewed on a rolling basis. We welcome researchers, builders, and funders with demonstrated expertise or strong interest in advancing AI-first approaches to security, privacy, and decentralized cooperation.