groningen-1452258.jpg

AI Safety Initiative Groningen

Our mission is to raise awareness of the full spectrum of existing and potential harms from AI, inform mitigation priorities through ongoing discourse, and support the realization of effective solutions.

We recognize that the rapid evolution of AI technologies presents both immense opportunities and significant challenges. As a community of forward-thinking individuals, we are dedicated to ensuring these technologies are developed and utilized in ways that uphold human dignity, foster equity, and prevent harm in both the present and future.

To achieve our mission, we focus on:

  1. Education: We provide educational resources, workshops, and seminars to equip our community with the knowledge and skills necessary to navigate the fields of AI safety, ethics, and alignment.

  2. Research: We support interdisciplinary research projects that explore the ethical, social, and technical aspects of AI safety, fostering collaboration between experts from various domains.

  3. Outreach: We engage with the general public to raise awareness about AI safety, ethics, and alignment, emphasizing its importance in shaping the future of technology.

  4. Networking: We facilitate connections between students, academics, and professionals who share our vision, fostering an environment of collaboration and innovation.

Together, we strive to create a world where artificial intelligence serves humanity and contributes to a prosperous, secure, and equitable future for all. Join us in our mission to shape the safe and responsible development of AI technologies, and make a lasting impact on the world.

Why AI Safety?

At AISIG, we stand at the crossroads of technological marvel and responsible guardianship. Artificial Intelligence (AI) is revolutionizing healthcare, education, and countless other sectors, yet it also holds the potential for unintended, or even catastrophic, consequences. From AI being exploited by malevolent actors and exacerbating social inequalities, to the more speculative, yet profoundly consequential risks of advanced AI systems seeking power autonomously, it is crucial to temper our excitement with vigilance. Our commitment at AISIG is to ensure AI benefits humanity sustainably and equitably.

Become Involved

At AISIG, fostering a community of knowledgeable and engaged AI Safety advocates is central to our mission. Through a diverse array of initiatives, we are creating an enriching environment for students, scholars, and enthusiasts from various academic backgrounds. Our offerings include AI Safety courses, hackathons, in-depth gatherings that facilitate critical discussions on AI Safety topics; speaker events featuring thought leaders; and an evolving array of exciting opportunities. Driven by a passionate team of organizers, AISIG is committed to equipping individuals to be informed contributors in the crucial field of AI Safety. Make sure to check out our events and become involved!

Join Our AI Safety Research & Discussion Groups

These groups cover topics ranging from mechanistic interpretability to AI’s impact on labor and democracy, and are meant for those keen to collaborate, gain hands-on experience, and find mentorship. By meeting regularly, they serve as a vibrant space for discussion, project work, and shared learning, guided by experienced professionals and/or students.

Selected research from our members

  • Steering Large Language Models using Conceptors (NeurIPS 2024)

    Joris Postmus and Steven Abreu [link]

  • Playing with Perception: Fooling Traffic Sign Classifiers via Copy-Paste Manipulation

    Davide Zani and Alexandru Dimofte [link]

  • Sandbagging LLMs using Activation Steering

    Jeremias Ferrao and Davide Zani [link]

  • Cybersecurity Persistance Benchmark

    Davide Zani and Jeremias Ferrao [link]

  • AI Misinformation and Threats to Democratic Rights

    Davide Zani, Mariam Ibrahim, Tiwai Mhundwa, Felix Michalak, and Andrie Avram [link]

  • AI Job Displacement and What Can Go Wrong

    Felix Michalak [link]