Über die Motivation von KI (Künstliche Intelligenz)

Video in TIB AV-Portal: Über die Motivation von KI (Künstliche Intelligenz)

Formal Metadata

Title
Über die Motivation von KI (Künstliche Intelligenz)
Alternative Title
Künstliche Intelligenz: Der Begriff ist nicht besonders intelligent. Lernende Algorithmen...trifft es vielleicht besser.
Title of Series
Author
License
CC Attribution 4.0 International:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Identifiers
Publisher
Release Date
2019
Language
German

Content Metadata

Subject Area
Abstract
"Artifical Intelligence", zu Deutsch "Künstliche Intelligenz" ist ein großer Begriff, der mit vielen Ängsten, Wünschen und sonstigen Projektionen überladen ist. Lernende Algorithmen dagegen hört sich harmlos an. Sind sie aber nicht, meint ein Ex-Borg und möchte das in einem Lightning Talk erläutern.
Loading...
Wechselseitige Information Algorithm Artificial intelligence
YES <Computer> Microsoft Algorithm Online chat
Physical law Robotics
Inference
da of crystal open source professional so excited artificial intelligence strong ki should simulate real human thinking weak ki only supports special ones application areas as far as clear we can only look at the weak ones because I do not have such a great mind I try it no longer strong cadres record so artificial intelligence the concept is not that's why I have the most intelligent said that actually are learning the algorithms that may be better, but in the cinema title or in a lightning Talk title then does not fit so well how does the algorithm learn? how certain patterns are interpreted by bringing someone with them she trains and this is possible, for example with pictures we look at us Now a few mailboxes in Germany look like that and in
England may look like that already very different and hydrant and
from the correct detection of a may already be off human life they can look like that, that's the way it is a 3d model or that is 100 years unchanged is the are pretty cool thing if the rusted looks like this and in the meadow stands this way or sometimes colorful or so and that would be the But 100 in the Vahr but not hit right because if the hydrants in the when ki does not recognize the hydrant properly what happens then she may park before and after something like that happens that's why they have an instant actually the English letterbox in seen a few hans seen in jerusalem yes that was the one here that is not a hydrant but a mailbox but one english standing in jerusalem and the brinkies really messed up like that because learning needs so teaching and the teachers are mostly people and Aldi are with you have it prejudices already are with prejudiced and now let's go to Facial recognition that now everything possible two expired why shows What kind of color does that have for us? get here and so on that is alexandria and casio court it a new yorker congressman no 30 of the technician a lot and she stated facial recognition algorithms are critically the algorithms are based namely only on assumptions and ki turn this into automated assumptions and further, when we did the Prejudices will not clear out the just automate them and I would have not supposed to take such a dark red because you can hardly see that there if you click on the whole red one Things also come left with more pages where they can read it and for that she was laughed at because it was are only objective mathematical algorithms what the ksk do but that's right and only they will algorithms of people trained and used and therefore applies to how learn car before garmisch in garwitz automate algorithms automotives So prejudices because she has just right
had the really dangerous thing about it but that algorithm is very very much faster very much more decisions based on prejudice can do as a people news from three years ago microsoft ended the day because they were inside 24 hours became a nazi who does not want to believe that first title of a priest hit everybody or also a fucking money center v3 and burning hell and hitlers reduce so the
succinct answer of microsoft was then Write your neural network on read to set was a terrible mistake I think so too Now it's getting a bit serious privacy who says I have nothing to hide in spite of being able to say that when I would have married someone now make an experiment unlock it Please now your phone and submit it does not matter who sits there and then we'll see privacy for important eighth or not so and now they face times in front of a ki as well as this jackpot of microsoft is unleashed on it and is now making systematic mistakes but not visible on the internet but decide if you want the next one get criticized whether they are thrown in a jail or similar because then the question is who that actually repairs and what we do we know a little bit further laws eg dsgv above still gone so far the eye is wide let us simple
make laws and isaac asimov has already in the 40s with the three set robot outstanding the work done in accordance with me want to passivate a bit The first law a ki may be human never kill or hurt inflict or allow by inactivity that a human harm is done the second law every ki must be the private data of every person with your own virtual life and love body and life protect unless she would be against it First law violated and the third law a ki must have its own security against viruses and state protect the trojan as long as this is not against the first or second law violates no further many thanks
Loading...
Feedback
hidden