Artificial Intelligence

When people think about artificial intelligence (AI), the first thing that comes into mind (besides apocalyptic thoughts) is "unemployment". There is an increasing concern about the reach of AI systems and the consequences that it will have in the job market. A Youtube channel called CGP Grey (which I strongly recommend you to subscribe to, since it explains a lot of interesting stuff, and not just AI) posted a video called Humans Need Not Apply four years ago, discussing the implications of AI in our daily lives.

NOTE: This frame contains a playlist I have created about AI, where I include other interesting videos that will be discussed later, as well as ted talks about the subject. Feel free to see some of them!

Based on this video, the purpose of this week's philosophy club is to discuss the following four aspects that are highlighted:

1. (0:00) Baxter and friends vs low skill jobs?

As the video starts pointing out, there is a key difference between the mechanical power that replaced physical labour in the industrial revolution and the new industry 4.0 revolution: intelligent automation. A robot arm in the assembly line of a car factory was designed by a human to perform one specific task, such as painting a car's door. However, thanks to the recent advances in AI it is now possible to design systems that can equally paint a door and then learn how to paint the rest of the car without specifically programming it to do so. 

Numerous concepts in the Machine Learning (ML) area have come into existence, such as Supervised Learning, Active Learning, Interactive Learning, Reinforcement Learning and Imitation Leaning (a 101 guide on the most basic ones can be found here). Each of these strategies serves a different purpose, but as the introduction of these video points out, their first target is to get rid of the "low skill" workforce. But is this a concern for the general public, or just a "logical" next step of automation? At this point, this only seems as a new way to cover low skill tasks, but this time using fancier machines. Take for example Baxter or EDI (video below), who can equally serve a coffee or fold your clothes by simply looking at the task. Obviously all of these mechanical tasks depend also on a good physical design of the robot (i.e. mechatronic, largely related but NOT the same as AI) and are largely confined to the a particular space. However, is this a first step towards a broader spectrum of tasks?


Further Reading:

• Low skilled workers to face disadvantages as robotics, AI take over: ADB. The Economic Times. https://economictimes.indiatimes.com/jobs/low-skilled-workers-to-face-disadvantage-as-robotics-ai-take-over-adb/articleshow/63715526.cms
• Industrie 4.0 From the Perspective of Applied Research. Prof. Reimund Neugebauer. Keynote Session of the IEEE International Conference on Emerging Technologies and Factory Automation. http://www.etfa2016.org/images/keynote/keynote-neugebauer.pdf

2. (5:00) Autonomous driving: Closer than you think. 

One thing is for sure: we NEED autonomous driving, and we will get it soon. Some people still think that it is dubious that we will live to see it happen, but just take a look at the streets of New York City 100 years ago, when cars where about to become a thing...

http://www.lifebuzz.com/street-life/

 

Back then sidewalks were optional, but in way less than 100 years, humanity was capable of transforming every major city in the world into "car-centric" spaces. Why do I say that we NEED this technology? Well, let CGP Gray explain it in a nutshell


So, are you now convinced that autonomous driving is the ultimate and unequivocal solution to all road problems? Some people still resist to the idea, such as truck drivers in Canada (video below at 3:55) and the media (apparently), who this year have reported two fatal accidents in the road provoked by autonomous driving cars. Also, some discussion has been sparked concerning a death algorithm, which is the concept of programming the car to take crucial decisions in saving lives/property in case of an accident.

3. (7:05) AI takes jobs... even AI jobs!

At the beginning of this section, CGP Grey explains that automation engineers are "skilled programmers whose entire job is to replace your job with a software bot". He also says that "the cutting edge of programming is not super smart programmers writing bots, it is super smart programmers writing bots that teach themselves how to do things the programmer could never teach them how to do". Last year, CGP Grey attempted to explain how this "algorithmic bots" works in this very illustrative video.

As these bots become more general purpose, is everyone's work at stake? Once that these bots know how to do a task, they can increasingly learn from experience until no further human interaction is needed. As a result, doctors, lawyers, stock brokers, and even AI researcher posts, are in peril.

4. (11:20) Be afraid too, you creative snowflake... 

The final barrier for AI is to take our creativity... o maybe it already did and we didn't notice. Some systems have already been capable of composing music, poetry and art. But is there anything else to conquer for AI? Are there still any loose ends in this story?