Document
Info︎    
                Welcome to Ruiqi’s archive table!                                                                                                                
week4 


The google AI Assistant reminds me of a news from Chinese social media recently.


A man got a call from the police asking if he had been scammed. The man later found that it was his AI Assistant who answered the phone and chatted with the swindler for three minutes. Then one netizen commented, "Are the swindlers stupid? Why can't they tell the voice of AI?" Another commented, "Maybe the swindlers are also AI.” This is a quite interesting example. Seems AI has been permeated into our every daily life.

In an age when AI can answer phone calls, what's the difference between an AI and a human?

“first is that they would never be able to use words or other signs by composing them as we do to declare our thoughts to others.” “For we can well conceive of a machine […] that emits words […] but it is not conceivable that it should put these words in different orders to correspond to the meaning of things said in its presence.”


It's true that machines can learn skills with a lot of training. Thinking about Do Androids Dream of Electric Sheep?, It seems that if ignoring physical differences, when machines also have emotions, they can be happy/angry/jealous/angry. In the dialogue between the AI and the fraudster, the AI seems to be flirting with the  fraudster. Does this indicate a new level of consciousness embedding in modern AI?


“One interesting definition has been proposed by A. M. Turing: a machine is termed capable of thinking if it can, under certain prescribed conditions, imitate a human being by answering questions sufficiently well to deceive a human questioner for a reasonable period of time. “


CAPTCHA reminds me of the human labourers who work on the image algorithms industry.
https://unthinking.photography/articles/unevenly-distributed


“People from Venezuelan made up to 75 percent of the workforce on some of the largest platforms specialised on crowdsourced image annotation in 2018 and 2019..... Humans have to, for example, draw so-called bounding boxes around cars or have to assign descriptive labels to every pixel in the video frame. These so-called semantic segmentation maps are currently the most common and most time consuming of various forms of image labelling.”


In fact, AI is behind a large number of cheap labourers. The training of AI image recognition is not generated in one day, nor is it through some coding system that can teach AI to recognize these images. It is more through human labor. Essentially, those big companies represent capital, consume and exploit cheap labors, and then build AI systems one after another. What people can see is the technological progress represented by ARTIFICIAL intelligence, but what they can't see is that there are actually human beings behind these technologies.
                                                                                                                                                  @Ruiqi Li 2021