Artificial General Intelligence
Course Information
- Organized by: Prof. Joschka Boedecker, Julien Brosseit
- HISinOne: 11LES-510-33 - Proseminar Artificial General Intelligence
- Kickoff meeting: Friday, April 24, 16:00-17:00, NEXUS Lab, IMBIT building (Georges-Köhler-Allee 201)
- Location: The seminar will be held en bloc, most likely shortly after lectures have ended. Details tba.
- Language: English
- Email:agi-seminar@informatik.uni-freiburg.de
Overview
With the remarkable success of large language models (LLMs) in recent years, along with their growing capabilities in handling multimodal inputs such as audio and images, improved memory, and advanced reasoning, the topic of Artificial General Intelligence (AGI) has gained increasing attention. AGI is often described as an intelligence comparable to that of humans. However, there is significant disagreement on what AGI actually is, how it should be defined, what its potential impact might be, and whether or when it will emerge.
In this seminar, we aim to explore these fundamental questions. Students will examine how AGI could be defined, measured, and evaluated, review current research directions, security aspects, and discuss the broader ethical and societal implications as well as future outlooks.
Format
This course is organized as a block seminar. It begins with an initial meeting that provides a detailed overview of the course. Following this, students will have the opportunity to vote for their preferred papers from a curated list. Based on these preferences, each student will be assigned a paper.
Students will work in teams of two, carefully study their assigned paper, and discuss it with their supervisor. They will then design a poster and present their work during a final poster session.
Timeline
| Date | Comment | |
| Introductory Meeting | ||
| Announcement Paper List | ||
| Paper Voting Deadline | ||
| First Meeting with Supervisor | ||
| Second Meeting with Supervisor | ||
| Final Poster Deadline | ||
| Poster Session |
(Preliminary) Paper List
- "A Formal Measure of Machine Intelligence" (Legg et al. 2006) [link]
- "A Definition of AGI" (Hendrycks et al. 2025) [link]
- "ARC-AGI-3: A New Challenge for Frontier Agentic Intelligence" (ARC Prize Foundation 2026) [link]
- "Understanding and Benchmarking Artificial Intelligence: OpenAI’s o3 Is Not AGI" (Pfister et al. 2025) [link]
Resources
Poster guideline as pptx or pdf