Michael Anissimov on Friendly AI

Michael is reviewing development in AI up to this point.

A lot of big developments in AI in the 2000’s and in the last few months.

Discusses the use of AI by corporations and military and others.
Provides another perspective on Hod Lipson’s work.

Reviewing the mainstream recognition of AI
Wired talked about Eureqa – download a scientist.
Robots performs Science (4th top scientific discovery of 2009)

Other AI successes
* air force funded research developed new algorithms which cut overhead for analysis of aerial surveillance photos by 99%
*Wolfram Alpha- 10 trillion pieces of data, 50,000 types of algorithms and models and linguistic capabilities for 1000+ languages
* University of Chicago system analyzes ultrasound images to determine whether breast cancer tumors have metasized
* University of Granada – document seach engine ase on AI for Spanish Parliament searches for portions of documents and videos
* Northwestern University- News at Seven, AI program generates convincing sports reporting videos with animated characters
* Marcus Hutter – Monte Carlo AIXI Approximation, universal reinforcement learning agent plays pac-man
*Continously improving voice recognition, facial object recog, and 3D mappoing technologies -download iPhone Dragon dictation app for free
* DARPA funded machine reading program (MRP) to develop AI that automatically translates natural text into formal representations

Conclusion – The AI Winter ended in the early 1990s and AI has been going strong since

Exploring human-equivalent AI and Beyond

Ray Solomonoff Paper
Milestone B – genereal theory of problem solving

Milestone C. A critical point in AI development would be a machine that could usefully work on the problem of self-improvement. Newell and Simon were not successful in their attempts to get their “General Problem Solver” to improve it’s own methods of operation. While Lenat’s “Eurisko” has been successful in several problem areas, he has not been able to get it to devise good heuristics for itself. He is, however, optimistic about the progress that has been made and is continuing this work.

Milestone D. Another milestone will be a computer that can read almost any English text and incorporate most of the material into its data base just as a human does. It would have to store the information in a form that is useful for solving whatever kinds of problems it is normally given.
Since there is an enormous amount of information available in electronic data bases all over the world, a machine with useful access to this information could grow very rapidly in its ability to solve problems and in a real sense in its understanding of the world.

Milestone E will be a machine that has a general problem solving capacity near that of a human, in the areas for which it has been designed — presumably in mathematics, science and industrial applications.

Milestone F will be a machine with a capacity near that of the computer science community.

Milestone G will be a machine with a capacity many times that of the computer science community.

Michael has a blog post on Solomonoff AI goals and speed of AI takeoff

Soon AI help make a robot that can walk out of the printer. Not a robot army but a roomba army

The AI advantage
* Faster
* Smarter
* Stronger
* Self replicating
* Be in many places at once
* Better learning procedures
* Photographic memory
* extensible hardware
* Access to own source code

Incorrect perceptions of AI
god complex issues
People anthropomorphise everything even Mr Potato head.
Even basic AI can kill (drones, robotic cannon accidently killed 9 people) thus the need to engineer friendliness

asimov laws were plot device

Basic AI drives Stephen Omohundro