Goals that hold
Alignment
Teaching capable systems to pursue what we mean — not merely what we say — so their objectives stay faithful as power scales.
Lisbon AI Safety Hub · Portugal
The most important work of our time is keeping AI safe.
We give Portugal a voice in the development of advanced AI by bringing together researchers, academics, policymakers and ambitious individuals. We aim to be a part of creating safe AI for all humanity.
What is AI safety?
Capability is compounding.The hard question is whether we can keepthat power under human direction.
Seeing inside
Interpretability
Making neural networks legible: reading what they represent so we can audit, predict failure, and intervene before it spreads.
Power with rules
Governance
Designing the norms, policies, and institutions that decide who builds frontier AI, under what limits, and with what accountability.
Programmes
However far along you are, there's a next step.
AI safety is a sociotechnical problem, so we support both governance and technical work. Pick the rung that fits today. Most people move through several over a term or two.
01Applications open August 1
Foundations of AI Safety Fellowship
A non-technical fellowship introducing AI risks and practical ways to help bring about safe AI for all humanity.
Explore the fellowship02Ongoing · open to allEvents & Lectures
Film screenings, directed discussion groups, and free-form meetings to build the community, plus lectures from industry experts.
See what's on03In person & onlineCareer Guidance
Get tailored career advice on opportunities in the field so you can make an impact.
Apply Now04Monthly · EN & PTNewsletter
Frontier AI news plus job and fellowship opportunities, delivered once a month.
Join the list05Community · always openDiscord server
Join the discord server to connect with community members and share resources and event information.
Join the DiscordFellowship
Foundations of AI Safety Fellowship.
Why AI safety matters
Understand the development of advanced AI, the risks it may create, and why this challenge matters now.
Understanding the risks
Explore the main ways advanced AI could affect society and how the field evaluates those risks.
Governance and action
Learn how policy, institutions, and international coordination can support safer AI development.
Your next step
Identify practical ways to contribute through your career, studies, research, policy, or community work.
Team
The people building LAISH.
We are a team building Lisbon's community for thoughtful, practical work on AI safety. Our programmes, events, research, and outreach help more people find a meaningful way to contribute.
This term
What's coming up. Events, lectures and socials.
Events held primarily in English unless otherwise stated. We will always have Portuguese speakers in attendance so we encourage everyone to join.
9SEP
LAISH Wine & Cheese Social
Wine and (vegan) cheese social to get to know the community and find out about upcoming events with LAISH.
WED 20:00Attend
23OCT
Film Screening: The Thinking Game
Film screening of The Thinking Game plus discussion and light refreshments.
FRI 20:00Attend
Monthly discussion circles
A roundtable on frontier AI research, once a month. Light refreshments served.
12OCT
Frontier Paper Discussion Group
Discussion of frontier research in AI in a roundtable format. Light refreshments served.
MON 19:00Attend
16NOV
Frontier Paper Discussion Group
Discussion of frontier research in AI in a roundtable format. Light refreshments served.
MON 19:00Attend
14DEC
Frontier Paper Discussion Group
Discussion of frontier research in AI in a roundtable format. Light refreshments served.
MON 19:00Attend