Beaverstate

Is artificial intelligence a worthy goal or a very dangerous idea.

I tend towards its a bad idea. Why would one (humans) want to create something potentially more intelligent than him/herself?

However we seem to be rushing straight ahead.

paddingtongreen

Why do people climb mountains, swim the English Channel, develop theories on Relativity, evolution, Quantum Mechanics, etc. You can't stop the AI developers, it is a challenge that must be met.

Bob

AI is great....they don't necessarily have to be that smart...just smart enough to clean the bog...

Yorkieabroad

Well that's smarter than my 10 yo.....he can't even hit it, never mind clean it!

lizzyq

It does get better! Have you managed to get him to lift the seat?

Yorkieabroad

The lid would be a start.....

GeoffM

I was watching something on NOVA about a computer contestant on Jeopardy. Fascinating how an entire team was training their racks of machines to be able to understand the questions. Some of the nuances of the English language means true AI is still a very long way off - for one, they need to combine different senses (sight and sound) to be able to read the non-verbal communication that we exhibit when talking. Even the Jeopardy computer had the questions fed to it electronically rather than by voice recognition.

I won't spoil the result, suffice to say it was extremely close!

steveq

I think we should move onto artificial intelligence when we've mastered the natural sort.

Ichigo-chan

All my intelligence is artificial

Speedwell

You mean, all your artifice is intelligent?

The only crack I know about artificial intelligence is the old joke about what you call it when a natural blonde dyes her hair brown...

Ichigo-chan

Touche!

Michael

AI is just a machines ability understand it's environment and to learn from it's mistakes. The best know and easiest understood use of AI is when a computer plays chess. Most computers play chess by a certain set of rules and will make the same mistake over and over again losing to humans (no AI) but a computer programed with AI may learn that specific moves will always produce the same results (a loss) and therefore will try different moves to try to win and if the move is successful, it will use that move in the future until that move fails and then it may try a different move always refining it's moves.

AI is not bad as long as the human programmer does not give the computer the ability to make a final decision in a critical situation such as launching missiles based on it's perception of the environment.

AI sometimes can have undesired results such as when used for high frequency trading. As an example during the last month, high frequency trading computers were monitoring messages from important sources when they saw a twitter message from the Associated Press indicating that two explosions occurred in the white house and Obama was injured and the computers started selling driving the DOW down 150 points for about 2 minutes until it was realized that someone had hacked into the Associated Press twitter feed. Some of the computer programs without very good AI lost a lot of money on a false rumor as humans bought when the price was low but computers with good AI may have actually made money since they sold quickly and decided that when the market again started to rise that the message must be wrong and started to buy back quickly making a profit on the false rumor.

Zen10

Not long ago, a plane on final approach to a runway in Scandinavia suddenly started a full climb with maximum thrust while letting the airspeed drop almost to stalling speed. The pilots had to turn the autopilot off and regain control of the aircraft before it crashed. They think it was a false glide slope reading by the computers. The computer had no way of knowing it was about to crash the aircraft because it thought it was doing the right thing. That's the difference between AI and human intelligence.

But this is one of those things whose inevitability renders our opinions somewhat irrelevant!

lansbury

What most Oregonians use when driving

steveq

What about the Air France crash, where the pilots completely misinterpreted the airspeed indicators, screwed up several other key things and stalled the plane and crashed. It doesn't take artificial intelligence to do something really dumb.