Laura

February 4th entry.

I found an article that talks about quite interesting aspects of our relationship with robots. It talked about why we allowed violence towards robots, and should it be considered as acceptable. The study concluded that the fact of not having empathie for the machines explain why violence is accepted: “The experience of walking in the shoes of a robot led the participants to adopt a friendlier attitude”.

Our treatment of robots also is influenced by our gender and racial bias. For example, black robots were more likely to experience violence than a white robot, reflecting our society.

“For now, faced only with non-sentient robots, we probably don’t need to worry about our bad behavior having deleterious effects on them.But some researchers do worry about it having deleterious effects on us, skewing our morality. If we hurl sexual abuse at female voice assistants like Alexa or direct racism at black robots — and get zero pushback because “it’s just a robot” — that could make us more inclined to mistreat actual women and people of color.”

However, some researchers predict that we may over-empathize with robots, and that we may prefer them than humans. They will be predictable and always in a good mood, contrarily to a human being.

I don’t know what to expect towards sentient robots; it still remains science-fiction in my head. But no matter what happens, the sociological side of the relationship we will have with these non-living beings will surely be very interesting.

February 16th entry

Infographic about “I’m Not Batman”, listing the types of intelligence needed to understand the story. However, the subject I would be interested to work on would be the ethics of AI in war. I heard about SKYNET in this article:

“A new examination of documents detailing the US National Security Agency’s SKYNET programme shows that SKYNET carries out mass surveillance of Pakistan’s mobile phone network and then uses a machine learning algorithm to score each of its 55 million users to rate their likelihood of being a terrorist.he data scientist Patrick Ball, director of research at the Human Rights Data Analysis Group, which produces scientifically defensible statistics about human rights abuses, called the NSA’s methods “ridiculously optimistic” because a flaw in how the NSA trains the algorithm to analyse cellular metadata makes the results unsound. Most of the 2,500 to 4,000 people killed by drone strikes since 2004 have been classified as “extremists” by the US government. They may in fact have been innocent.”

March 4th entry

My infographic about predictive analytics, taking the example of Netflix (second one). It is linked to the one of Maria (top one), which is a descriptive infographic of predictive analytic.

March 11th entry

In short: Predictive analytics is about using data in order to make predictions. It is often used to predict customer habits and to recommend products or services. For example, Netflix uses their customers data surrounding scrolling behaviours, searches, where you pause, rewind or skip ahead, and so on, in order to personalize the recommendation system. This is why the content suggested to you is very different from someone else, and can seem quite homogenous. Significantly, the recommendation system influences 80% of what we watch on Netflix.