Growing the AI Safety Community in Portugal

Who we are:

The Lisbon AI Safety Hub (LAISH) is the first, and thus far only, AI Safety community in Portugal. We are building Portugal’s capacity to address transformative AI risks through fellowships, career support and community events. We are part of the growing landscape of AI safety communities in Europe working on growing the field and contributing to the development of safe AI. We are distinct in this landscape because we focus on AI Safety efforts where Portugal and the EU can have the most comparatively meaningful impact (e.g. extreme concentration of power, gradual disempowerment, compute governance).

Here is a list of activities we are putting on and planning:

8-Week Introduction to AI Safety Fellowship: The foundations of understanding AI safety and governance, based on coursework developed in partnership with the team at Lens Academy. We also have guest lecturers planned to give context on field-building and opportunities as a part of the fellowship. We currently have 30 fellows across 2 cohorts with the first two sessions resulting in 90% attendance rate. Fellowship attendees have already become contributing members of our organization. 2 are running our events, 4 are working on a government outreach and lobbying project and 2 are writing for our Substack. 4 of our fellows will be attending EAGx Berlin after learning about it through our community.

Discussion Groups: Monthly discussion circles talking about frontier research and policy papers for members of the community. Venue space has been secured for the first three monthly sessions.

Film Screenings: One screening of an AI safety related film every two months, open to the public, followed by group discussions led by the LAISH team. The first scheduled event had 50 sign-ups and was held on October 23rd. The second event will be held in December and the third is tentatively planned for February.

Discord Channel: Growing community following along with our communications where we post opportunities, new research, AI safety news and upcoming activities from LAISH.

Substack Newsletter: Monthly newsletter with frontier research and news. We also publish articles from our team members and fellowship participants on AI safety related topics.

Community Events: Bi-weekly meet-ups, hang outs and board game nights, jubilee style debates among other things, to build the community and create deep ties.

Local and International Partnerships: Bringing a perspective on catastrophic risks to local panel discussions and workshops with organizations here in Lisbon. Also partnering with groups around the EU, particularly in Madrid and Barcelona to put on joint events like hackathons.

Team:

Leadership:

Caleb Eli Price****: Co-founder and president of LAISH. ML4Good, Human Aligned AI Summer School, Gen Stream alumnus.

João Bexiga****: Co-founder and Head of Operations. Former CEO of IT consulting and digital services firm Aubay.

Core Team:

Orlando Torres: Chief of Staff. Founder of 4 different EA organizations.

Tânia Girão: Head of Communications. ML4Good alumnus, Volt board member.

Frederico Luz****: Discussion Groups Coordinator. Digital Technology and Artificial Intelligence scholar.

Mateus Carvalho: Government Outreach Campaign Manager. Center of Effective Altruism Onsite Coordinator, Volt Portugal Co-founder.

Volunteers:

Haimi Tefera: NOVA University Affiliate

Alice Parrinha: Tecnico University Affiliate

Advisors:

Julien Sireau - Operations Manager at Oxford AI Safety Initiative, Gen Stream Co-Founder and Founding President of AI Safety Cardiff University

Yaqoob Ahmad - Founder and Co-Director of Gen Stream. Operations associate at Impact Ops. Field builder and Generalist in AI safety.

Manon Kempermann - Generator resident, Pathfinder fellow and Program Director at Safe AI Germany

Tzu Kit Chan - Outgoing Chief of Staff at Atlas Computing, incoming Head of Operations at Robocurve

Sam Baskeyfield - Operations Associate at The Future Society

Alexander Müller - Director of Safe AI Netherlands (SAIN)

LAISH has grown extremely rapidly from just a single founder to become a full-fledged team with distinct roles and responsibilities. Our community has similarly increased extremely quickly with nearly two hundred members across our activities and fellowships. We foresee continued growth as the reception and outreach from the Portuguese community has been much stronger than expected. Funding would allow us to expand on and continue our operations into next year and reach more people and help more people transition into AI safety roles. Funding would pay for full and part time employment to run the organization, paying for a full-time president, and two part time staff to coordinate and run events and manage the fellowship.

Funding also allows us to have dedicated space for our events, provide catering for the events and pay legal and administrative costs to run the organization.

All of this is in service of capitalizing on the momentum we have gained in creating the organization and helping it to expand and capture more people. We want to reach as many people as possible and help them transition into AI safety roles or at the very least become literate on AI safety issues and catastrophic risks. We believe this will have a substantial impact on mitigating the negative outcomes from transformative AI development.

Why LAISH is impactful and important

• Portugal is an EU Member-State. All EU countries have disproportionately high leverage in affecting policy because of their ability to propose legislation at the EU level, namely through national government’s voting power in the European Council and Council of the European Union, as well as talk with and influence other EU politicians and get international attention about their AI policies. An example of concrete, dedicated action on AI safety is the EU AI Act,  the first comprehensive regulation for AI anywhere in the world, which shows that the EU plays a role in AI safety.
• Portugal had no AI safety hub before LAISH. We fill that gap as, besides us, there is no capacity for those living in Portugal and interested in AI safety to get involved, besides moving away or joining foreign organizations. This is not always an option for everyone, nor should be the only way for someone in Portugal to be able to engage in AI safety.
• Local communities help to build the field from the bottom up in a grassroots manner. People who would not have been introduced to AI safety otherwise have a group that they can go to for information and advice. They can communicate with people in their local language and be inspired to study more and eventually take on AI safety roles.
• Hubs provide a physical space and foster a community of people to run fellowships and classes, share ideas, and cooperate on research and projects.
• Portugal has a strong expat community of remote workers and digital nomads who work in tech and might be interested in AI safety.
• Young Portuguese people are among the most highly educated in Europe, with 43% having tertiary education.
• Lisbon has 7 accredited universities and over 25 other higher education institutions. There is a lot of space to find researchers and partner with universities. We have already received significant interest from professors at two universities (ISCTE and Nova).
• Lisbon has been hosting Web Summit for the last 10 years, growing a foothold in the tech industry and bringing world-recognised leaders in the tech industry for years.

Funding Received So Far

BlueDot Rapid Grant - $6100 to begin running the fellowship and hosting events and meet-ups, to provide catering and event venue rentals as well as logistical costs.

What We Are Asking For

Minimum Funding: €62,000 (67,000 USD):

• Staff, 1.0 FTE, including employer social security and the 14 annual salary payments required under Portuguese law: €44,261
• Legal incorporation and accounting: €5,200
• Programme (fellowship, film screenings, guest lectures, community events and speaker travel): €6,080
• Event space (Impact Hub Lisbon membership): €2,400
• Software, supplies and website: €1,066

Maximum Funding: €115,000 (125,000 USD):

• Staff, 2.0 FTE, including employer social security and the 14 annual salary payments: €88,523
• Legal incorporation, accounting and a reserve for professional services: €10,900
• Programme (fellowship, film screenings, guest lectures, community events and speaker travel): €6,080
• Event space (Impact Hub Lisbon membership): €2,400
• Software, supplies and website: €1,414

A full breakdown of expenses can be found here.

Theory of Change: Our theory of change is centered around the idea of moving people up the ladder of AI safety literacy and context which causes more people to get involved in AI safety roles and thus make impactful differences through their work or political outreach.

The full theory of change can be found here.

A large local AI safety community provides a first point of contact and ways to get involved for people who are interested in AI safety but don't have much context ➡️ Some of those who get involved with the community will take our fellowship or find other ways to become highly knowledgeable on AI safety ➡️ They will transition into AI safety careers, grassroots organizing and government lobbying ➡️ Community members and course fellows will engage with politicians and convince them to adopt stronger AI safety policies ➡️ Portuguese politicians will vote on AI safety policies in the EU Parliament and Council and influence other European politicians to take a stronger stance of AI safety ➡️ International agreements to reduce catastrophic risk are more likely to occur with highly literate and motivated politicians in the EU.