Vision

The development of advanced AI poses existential risk to humanity, akin to the risks of nuclear weapons, pandemics and climate change, if not greater. Current development of AI is insufficiently prioritizing safety, and the lives of everyone on Earth are being gambled with by just a few politicians and businessmen who hold the majority of the power. This is an unfair and dangerous state of the world that needs to be rectified.

To give Portugal a voice in the development of advanced AI so that it is done safely, aligned to humans and is equitable. We are part of a larger global effort of AI safety communities working on public awareness, technical research and international policy and governance. We aim to address the broad range of societal harms from transformative AI, from job loss and deepfakes to catastrophic loss of life and extinction.

Portugal is home to a highly educated youth population, a thriving tech sector, and a large community of expats, digital nomads, remote workers and immigrants who have chosen Portugal for its good weather and lifestyle. Portugal is also a member of the EU and a close ally of the UK and other Western nations.

The latent talent and curiosity is strong, and the potential to make a difference is there, but without a community serving as a unified voice, not much can be done. LAISH is the community that brings together the various groups in Portugal from all over the country to work on AI safety problems and campaign for a more just future.

We aim to make AI safety a priority in Portugal. We aim to educate people on the risks and capabilities of AI, steer them towards work or volunteering for organizations that work on AI safety problems, encourage them to do research on AI safety topics, and advocate and spread the message so that AI safety is a concern among all citizens. We believe that every nation can do this and that it can become a worldwide movement. We also believe that EU member states have an outsized impact due to their regulatory power, and that ideas which start on a national level can spread to the EU through policy proposals and discussions among politicians and parliamentarians.

We realize the above changes through the following actions:

• Community events: facilitating dialogue on AI safety and getting people living in Portugal together
• Courses: structured courses to teach people the fundamentals of AI safety
• Career advising: individualized feedback and guidance to help people transition to AI safety roles
• International connections: connecting the landscape of AI safety communities in Europe
• Outreach: attending and speaking at external events and around universities to spread awareness
• Talent pipeline: accelerating people through the talent pipeline (see below)

LAISH sees people as existing somewhere on the pipeline. This has been adapted largely from Safe AI Netherlands (SAIN), whose vision we believe is a solid model to emulate.

At LAISH we focus on those in levels 0, 1 and 2 and try to accelerate them to the next step. For example, a 0 can be converted into a 1 if they attend one of our meet-ups and read our newsletter. A 2 can be converted into a 3 after completing our Foundations of AI Safety Fellowship.

LAISH is still a young organization with a lot of room to grow. We have many ideas about future projects and ways to grow and make a greater impact:

• Expanding our courses: We want to offer places to even more participants next semester and expand to an extended fellowship as well as an intensive one-week course.
• International hackathons: We plan to put on joint events with the AI safety communities in Barcelona and Madrid to give people something tangible to work on and facilitate collaboration in the Iberian Peninsula.
• Political outreach: We plan to start a campaign to lobby Portuguese politicians to sign commitments to the development of safe AI.
• Permanent location: We aim to have a permanent location where we can facilitate research collaboration and hold more events.