ENG – V.C. – OP OVERLORD 2 – ARTIFICIAL INTELLIGENCE ISSUES

By -

IF YOU HAVEN’T READ THE PREVIOUS PART, CLICK HERE.

OP OVERLORD

OP OVERLORD
ARTIFICIAL INTELLIGENCE ISSUES

AI AS TECHNOLOGY NEEDS TO GROWN IN ITS DECISION MAKING CAPABILITIES. POINT IS, DECISION MAKING REQUIRES A LOT OF ETHICAL CONSIDERATION. HOW DO YOU CODE THEM? ARE ETHICAL CONSIDERATION EQUAL ALL AROUND THE GLOBE? WHAT IF WE HAVE AI WITH DIFFERENT ETHICAL STANDARDS (just like we have human with different ethical standards)?

Now, I want you to focus on the following stuff, because it’s our target.

Artificial Intelligence Issues.

AI has to face:

 

Short TermLong Term
AI – ObjectStructural Unemployment

 

Algorithms fairness

 

Machine Ethics

 

Autonomous weapons

Determining human values for AI

Status of mankind

 

Controlling AGI&ASI

 

 

AI – SubjectLegal Status of AI

 

 

Well-being of AIs

 

Mind uploads

 

 

We can see Articial Intelligence Ethical Issues from 4 different points: first, as IA as an Object or as Subject and both of them in the short term and in the medium-long term.

 

Where we face AI as an Object, we see in the short term the following:

  • Structural Unemployment
    • This condition is the aftermath of massive atomization of tasks and their automation. Manufacturing industries will require less and less people. In a growing-demographic contest, where will it take us? More people, less jobs. It’s something we need to deal with.
  • Algorithms fairness
    • It’s one of the most interesting and important aspects for regulators and it’s a huge concern today.
      Think about algorithms that can learn from existing situations (it should be their basic pattern) and those situations have flows or are distorted. Think about MS AI on Twitter that in less than 24h turned from a very cute school girl to a devoted nazi who blamed Jews for pretty much everything. Image a company who has an AI that decides raises and bonuses based on productivity but has, at the same time, very generous maternity conditions. Women will be less productive. They won’t get any bonus. Fairness in algorithms has to do with the set of data (and data can lead to a flawed conclusion) they use and HOW they are related one with another.
    • http://fairness.haverford.edu/

  • Machine ethics
    • Can a machine suffer or considered capable of (something)? Should it possess human rights?
  • Autonomous weapons
    • Drones, robots that kill suspects in USA and so one. That’s lame.

 

 

Where, as we face AI as a subject, we see in the short term the following:

  • Legal Status of AI
    • What are they?

In the Long term, as we consider AI as an object:

  • Determining human values for AI
    • This is huge. We need to give AI some values to start with, but values are a product of our culture. And our culture is waaaay different from our neighbours’. Think about the differences we can have with Thai culture. How many differences can you spot in our culture of the XXI century looking from a western culture point of view of “just” 70 years ago? Think about women and the ad you can now find on the net and that nowadays would have been considered (rightly) misoginist. How are we going to change in the next 50 years?
    • Values have changed over and over, over the millenia.
    • We may set them with advanced values we now think they are (will be) better for us, so that AI will be morally superior.
  • Status of mankind
  • Controlling AGI and ASI

In the Long term, considering AI as a subject:

  • Well-being of AIs
  • Mind uploads
    • What if you can store yourself on a HDD? What if you can upload yourself into something mechanical or digital?

Such a feeling of pride when you see good manners... further readings

1 Comment to ENG – V.C. – OP OVERLORD 2 – ARTIFICIAL INTELLIGENCE ISSUES

    Lascia un commento

    Il tuo indirizzo email non sarà pubblicato.