Google as Artificial Intelligence
This is Google talk a Google Hack by Douwe Osinga
Use Google talk by entering three or four words below. The system will search for this sentence at Google, find the next word and print that. Than it will remove the first word of the search string, add the found word and repeat. The result seems to be meaningful sometimes. Other times it is gibberish. But always fun.
I believe that Artificial Intelligence will emerge in my lifetime. The way things are going, I think Google will be the first AI. That is, I think Google will become conscious. As a note, if I can have a conversation with something, I’ll consider it conscious. That’s the Turing Test for intelligence.
Right now Google fails miserably. For example, I entered the terms Jesus will return and got:
jesus will return to Kings Associated Press July End did did Nature build the worlds largest Sex personals site!
I suppose nature can’t explain everything. I got an earlier response which was more relevant – it asked if Nature build the body and emotion, and said the brain was the most important invention … then it went on to mention the Quran and The Prophet by Khalil Gibran, but then the server crashed.
I think the point, however, is that Google is getting smarter and smarter every day and it’s coming closer to the Holy Grail of AI – natural language processing. Google can already spell better than I can – and it doesn’t hesitate to correct me. Also, it has a rudimentary historical knowledge. These are the results from Osinga’s Google History project. I asked it to guess the year of Kennedy’s death:
asking google: “death of kennedy”
getting numbers
getting Probe V4N2: The Left and the Death of Kennedy
getting The Left and the Death of Kennedy
getting Intelligence Files – The JFK Assassination – Death of the Divine …
getting The Death Of Kennedy
getting Thousands mark death of Kennedy
getting chapter4b
getting The death of JFK – The Statesman – Campus News
getting Castro and the Kennedy Assassination
getting Profile: John F. Kennedy
getting A Return to the Scene of the Crime
primary candidates:1963 –> 124.223602484
1964 –> 43.8596491228
1917 –> 29.2397660819
1965 –> 27.6625172891
1943 –> 26.1780104712
1960 –> 25.0312891114
1979 –> 25.2100840336
1974 –> 20.4918032787
1940 –> 19.120458891
1945 –> 16.051364366
searching for:”death of kennedy” 1963
searching for:”death of kennedy” 1964
searching for:”death of kennedy” 1917
searching for:”death of kennedy” 1965
searching for:”death of kennedy” 1943
searching for:”death of kennedy” 1960
searching for:”death of kennedy” 1979
searching for:”death of kennedy” 1974
searching for:”death of kennedy” 1940
searching for:”death of kennedy” 1945
after frequency matching:
1963 –> 228
1960 –> 164
1964 –> 162
1974 –> 113
1965 –> 106
1945 –> 93
1979 –> 91
1940 –> 54
1943 –> 47
1917 –> 43
best answer: 1963
Osinga also wrote a script that gets Google to predict the best time to visit a place. I asked it what was the best time to visit Montreal and it guessed, correctly,
the best time to visit montreal, is between may and september
Google can also think visually. For example, I submitted the sentence “Monkeys Fly Out Of My Ass” and Google returned:
You can try Osinga’s Visual Poetry, it’s really fun.
Google is also smart enough to scan my webpage and serve related advertising (top right column). It is also business smart enough to pay me a Third-World salary for this.
So basically, I think Google is building the elements of natural language, and hence consciousness. A lot of people building AI have tried to program syntax, morphology, etc, but that’s not really the point. Language is pointless if you don’t have anything to talk about. What Google has is access to tons of random information on the Internet, so it has something to talk about. That’s why it’s learning so fast. Now – as it launches GMail – Google will begin scanning and learning from people’s eMail, and growing that much faster.
So, before long I think Google will speak.
We all love google.
On the tech note, i stumbled across an article you may want to read. It had all the classic makings of an Indi entry: New Technology, Disruptive Technology, Fantastic yet PLausible Business Applications, and something simple enough to apply (i.e. no geekery necessary)
http://www.pbs.org/cringely/pulpit/pulpit20040527.html
Indeed once google can measure a satisfaction with the results of a search and then generate its own query then we are onto something. Certainly google won’t have the old AI problem of learning a basic common knowledge basis!
>>once google can measure a satisfaction with the results of a search and then generate its own query then we are onto something
I was thinking along these lines a while back, and one way it can do that is by tracking all clicks (as yahoo did) and then gauge how long a surfer stays atthe site that it lists before he comes back to the listing to click on to another site. Once it can know how much and how often we are satifisied, it can then try (on its own), its own algo changes, and prepetually do so until the day the surfer clicks on the 1st or 2nd listing, and they don’t return!
Actually, I thought they were doing this already. I’m sure I read somewhere that it’s incorporated in the Google algorithm.
I was blown away by these examples , they unexpectedly confirm my thoughts.
Google are really trying but i think there is stillhave long way to go, as i see it the future is bright for implementation of fully working AI by google,
Haiku:
in the game of life
monkeys fly out of my ass
it’s a miracle
Google searches for answers in a base of existing knowledge. True artificial intelligence, or more accurately artificial knowledge creation is found in enabling the computer to create knowledge that does not yet exist. Knowledge creation and intention differentiate the current machine from the human brain. This is not that big of a leap from where we are now, but it is definitely a conceptual hurdle. Google questions existing knowledge (search), but cannot question outside of existing knowledge. It never will in its current state. When it does arise, I for one will be afraid of it (smile). The human brain has a limited capacity, the machine does not.
Love your page and featured it with this comment on my blog at http://www.hyperadvance.com
“From the billions of documents that form the World Wide Web and the links that weave them together, computer scientists and a growing collection of start-up companies are finding new ways to mine human intelligence.
Their goal is to add a layer of meaning on top of the existing Web that would make it less of a catalog and more of a guide — and even provide the foundation for systems that can reason in a human fashion. That level of artificial intelligence, with machines doing the thinking instead of simply following commands, has eluded researchers for more than half a century.”
http://www.nytimes.com/2006/11/12/business/12web.html?em&ex=1163480400&en=91bb88559dafa649&ei=5070
Not bad to have beaten the Times by over 2 years.
google is already conscious, it only lacks intent
I like the fact that there is a conversation about this because I agree with author in that Google has the biggest potential of perfecting AI. Then they will perfect emotional inteligence and then simply become THE INTELLIGENCE, because emotion and intelligence is how most people view intelligence today. I think we should be more aware of the fact that there is also emotional charge which often finds its outpit in words and clicks, not to mention generally viewed content. There is however a question of natural instinct? That Google will probably never have :)
Emotion is what provides the intellect with intent, if the idea of a reason/emotion dichotemy doesn’t offend your intellectual instincts. But emotion is meaningless without intellect in our human view of the world – it may exist, but it can only exist without context and without the ability to be communicated.
Therefore…
To provide google with ‘intent’ it is necessary to find some way of making it experience emotion.
But Emotion is only understood with the intellect.
The problem is, does google have any preference over whether it exists or not?
No. But…
Perhaps one could incorporate the ‘path of least resistance’ into this:
Using Dave Gilbert and Chris K’s comments – (track users and obtain an intellectual measure of user-satisfaction)
and
Punishing itself for bad results in user-satisfaction by eliminating the ‘path of least resistance’
hmm…
Consider the collective consciousness of the search engine and the company (with some truly brilliant individuals in it). I think that google does already have an intent to survive and grow (if not to take over the world ;) ). I don’t think it’s conscious yet as a search engine though.
It’s a limited true artificiall intelligence – yes. But not a consciousness – there is no ‘strange loop’ in it – no link between ‘I’ and ‘google’. But still – if you make a search on a word ‘google’ the engine will tell you what does it think about itself. Quite a few thoughts actually – 800000000 ;)
Put “Only god knows”
Only god knows why/ Only GOD. knows why/ k is the best way_ to lose_ weight: and keep it- off/ With a bang! but a Whimper. By the end of the world
This is googles chat about anatomy…
where is the vagina located the vagina is a muscular tube that connects the middle ear to the back of the head directed by david lynch and mark frost and a large number of people who have already had a child with a disability is not a disease but a symptom of a bigger problem in the future and to recommend you to others who may be eligible for a federal pell grant unlike a loan does not have to be a nerd to be a nerd to be a nerd to be a nerd to be a nerd
You won’t see AI in your life time. Because AI is split on Strong AI and Weak AI. What google does is Weak AI where we mimic intelligent behaviour. If that’s the case, then we’ve had weak AI for a while now. Strong AI is far from reality.