AI and AI Safety: Generalization and AI Safety Research (Part 2)
Explore “Generalization and AI Safety Research” in this AI and AI Safety lesson. Part 2 of the five-part series brings together the source material and lesson topics. Use the contents links to find a section, read the notes, and consult the complete original teaching slides below.
Lesson topics
- There are a number of different forms of learning as applied to artificial intelligence.
- More challenging is the problem of implementing what is called generalization.
- WHY RESEARCH AI SAFETY?
- As pointed out by I.J. Good in 1965, designing smarter AI systems is itself a cognitive task.
Contents
- Read the original document
- Psychologists generally do not characterize human intelligence by just one trait but by the combination o
- There are a number of different forms of learning as applied to artificial intelligence.
- More challenging is the problem of implementing what is called generalization.
- WHY RESEARCH AI SAFETY?
- As pointed out by I.J. Good in 1965, designing smarter AI systems is itself a cognitive task.
REQUEST THE COMPLETE DOCUMENT ON WHATSAPP
Read the original source document
Artificial Intelligence – Gayathry Satheesan: original teaching slides for this study series.
Psychologists generally do not characterize human intelligence by just one trait but by the combination o
Psychologists generally do not characterize human intelligence by just one trait but by the combination of many diverse abilities. Research in AI has focused chiefly on the following components of intelligence: learning, reasoning, problem solving, perception, and using language.
There are a number of different forms of learning as applied to artificial intelligence.
There are a number of different forms of learning as applied to artificial intelligence. The simplest is learning by trial and error. For example, a simple computer program for solving mate-in-one chess problems might try moves at random until mate is found. The program might then store the solution with the position so that the next time the computer encountered the same position it would recall the solution.
More challenging is the problem of implementing what is called generalization.
This simple memorizing of individual items and procedures—known as rote learning—is relatively easy to implement on a computer. More challenging is the problem of implementing what is called generalization. Generalization involves applying past experience to analogous new situations.
WHY RESEARCH AI SAFETY?
WHY RESEARCH AI SAFETY? In the near term, the goal of keeping AI’s impact on society beneficial motivates research in many areas, from economics and law to technical topics such as verification, validity, security and control.
Whereas it may be little more than a minor nuisance if your laptop crashes or gets hacked, it becomes all the more important that an AI system does what you want it to do if it controls your car, your airplane, your pacemaker, your automated trading system or your power grid.
As pointed out by I.J. Good in 1965, designing smarter AI systems is itself a cognitive task.
Another short-term challenge is preventing a devastating arms race in lethal autonomous weapons. In the long term, an important question is what will happen if the quest for strong AI succeeds and an AI system becomes better than humans at all cognitive tasks. As pointed out by I.J. Good in 1965, designing smarter AI systems is itself a cognitive task.
Such a system could potentially undergo recursive self-improvement, triggering an intelligence explosion leaving human intellect far behind.
Review questions and answers
What are the key points about Psychologists generally do not characterize human intelligence by just one trait but by the combination o?
Psychologists generally do not characterize human intelligence by just one trait but by the combination of many diverse abilities. Research in AI has focused chiefly on the following components of intelligence: learning, reasoning, problem solving, perception, and using language.
What are the key points about There are a number of different forms of learning as applied to artificial intelligence.?
There are a number of different forms of learning as applied to artificial intelligence. The simplest is learning by trial and error. For example, a simple computer program for solving mate-in-one chess problems might try moves at random until mate is found. The program might then store the solution with the position so that the next time the computer…
What are the key points about More challenging is the problem of implementing what is called generalization.?
This simple memorizing of individual items and procedures—known as rote learning—is relatively easy to implement on a computer. More challenging is the problem of implementing what is called generalization. Generalization involves applying past experience to analogous new situations.
What are the key points about WHY RESEARCH AI SAFETY??
WHY RESEARCH AI SAFETY? In the near term, the goal of keeping AI’s impact on society beneficial motivates research in many areas, from economics and law to technical topics such as verification, validity, security and control. Whereas it may be little more than a minor nuisance if your laptop crashes or gets hacked, it becomes all the more important that an…
What are the key points about As pointed out by I.J. Good in 1965, designing smarter AI systems is itself a cognitive task.?
Another short-term challenge is preventing a devastating arms race in lethal autonomous weapons. In the long term, an important question is what will happen if the quest for strong AI succeeds and an AI system becomes better than humans at all cognitive tasks. As pointed out by I.J. Good in 1965, designing smarter AI systems is itself a cognitive task. Such…
More lessons from this document
- Artificial Intelligence – Gayathry Satheesan — Part 1
- Artificial Intelligence – Gayathry Satheesan — Part 2
- Artificial Intelligence – Gayathry Satheesan — Part 3
- Artificial Intelligence – Gayathry Satheesan — Part 4
- Artificial Intelligence – Gayathry Satheesan — Part 5
Conclusion
The important points in WHY RESEARCH AI SAFETY? should be reviewed alongside the definitions, examples and comparisons given above. Continue with the other four articles to cover the rest of the source material.