Administrative Information
| Title | Generalizability and Artificial General Intelligence (AGI) |
| Duration | 45-60 |
| Module | C |
| Lesson Type | Lecture |
| Focus | Technical - Future AI |
| Topic | Open Problems and Challenges |
Keywords
AGI,Generalizability,LLMs,Transformers,
Learning Goals
- Understand the limitations of currrent AI approaches
- Learn the definition of Artificial General Intelligence (AGI)
- Get familiar with the capabilities and core requirements of AGI
- See how we can test for AGI
- See how far away is AGI and understand the benefits and risks
Expected Preparation
Learning Events to be Completed Before
Obligatory for Students
- Introduction to machine learning and deep learning concepts given in previous lectures
Optional for Students
None.
References and background for students
None.
Lesson materials
Instructions for Teachers
The goal of this lecture is to provide students with an introduction to the idea of Artificial General Intelligence (AGI). It should set the stage for more in-depth discussions and debates about AGI. The lecture should:
- Clarify the differences between the levels of AI
- Discuss the expected characteristics of AGI
- Outline the limitations of the current state-of-the-art AI in terms of the characterists of AGI
- Present possible ways in which we might test for AGI
- Look at various expert viewpoints on how far away AGI is
- Discuss the possible benefits and risks of AGI in terms of Human Centered AI
Outline of the lecture
| Duration | Description | Concepts | Activity | Material |
|---|---|---|---|---|
| 10 min | Limitation of current AI approaches | Reliance on data and teaching (learning from limited data), Human scale nerual networks, offline learning versus continuous learning and adaptation of beliefs, integration into a complete AI stack | Taught session and examples | Lecture materials |
| 5 min | Definition of Artificial General Intelligence (AGI) | How can we define AGI, levels of AI (weak, strong, super) | Taught session and examples | Lecture materials |
| 10 min | Capabilities and core requirements of AGI | Sensory perception, motor skills, natural language understanding, knowledge retention, problem solving, common sense, creativity, consciousness, pattern recognition versus modeling the world | Taught session and examples | Lecture materials |
| 5 min | How can we test for AGI? | AGI Turing test, Coffee test, Robot college student, Employment test | Taught session and examples | Lecture materials |
| 5 min | How far away is AGI and what are the benefits and risks? | Metrics (time, speed of technological advancement, singularity breakthrough), Expert views and predictions, Possible outcomes and ethical concerns (Utopia, Status Quo, Distopia) | Taught session and examples | Lecture materials |
| 5 min | Conclusion, questions and answers | Summary | Conclusions | Lecture materials |
Acknowledgements
The Human-Centered AI Masters programme was Co-Financed by the Connecting Europe Facility of the European Union Under Grant №CEF-TC-2020-1 Digital Skills 2020-EU-IA-0068.