Artificial Intelligence, The Search For The Perfect Slave

Artificial Intelligence, IA, is a fantastic scientific and engineering project whose main purpose is to create a perfect slave. It may sound creepy or disrespecful, but the dream is to make a device clever enough so that one can acquire it, get it out of the box, swiftly ensemble it and ask it “get me a good cup of coffee”, just as the best human assistant, recently hired, would do. Of course, the test would be that one can enjoy a delicious coffee within a reasonable time.

The term Artificial Intelligence was coined in 1955 in Dartmouth, USA, by Prof. John McCarthy (McCarthy et al, 1955), to define a project on “the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it”. That project flew from the hands of McCarthy and others to become a huge, global and technological effort, which is often characterized in 2 ways: 1.- It is an engineering project aimed at designing (and building) one (artificial) device as smart as a human being. 2.- It is the flagship project of cognitive sciences to explore and characterize intelligence (of any kind).

This is a fairly simple (and innocent) characterization of the scientific-technological project that has occupied a community of thousands of researchers around the world for 60 years, with huge financial investments. A failed project, it seems, because there is no known device as smart as a human being, as yet.

The main problem to establish success or failure of IA is that we may not know everything. How are we going to establish whether there is even a small improvement on these complex endeavors if we cannot see the code when it comes to software?. Hiding code is always a personal, presumably egotistical decision, probably prompted by some alleged need to preserve strategic advantage over competitors. Working within that mindset is like living in a hostile jungle sourrounded by predators. So, one should expect some people the pressure to hide intelligent code. But, to develop complex systems, that may not be a very smart strategy, as the support of a community may make all the difference to a healthy, creative and robust development.

This is why we welcome FLOSAI: Free, Libre, Open Source AI. It exists and it is getting more and more succesful and enticing. You do not have to take our word for it, of course, as being open means that everyone can see it all for themselves. Let us just point to three examples of good FLOSAI, with different levels of development and technical orientation.

In realistic terms, looking for the perfect slave does not have to resemble the very sad history of human slavery. These new slaves do not have to be badly treated and exploited. Just the opposite. They are treated as friends, if only because they represent the efforts of many intelligent humans to deal with loading and engaging tasks.

Take, as the first example, Mycroft, which started as “a friendly AI virtual assistant for Linux users” but, as you will see, can find its “places” at home to help. And, it is open source. So, there is no limit to what you can adapt it to.

To show that intelligence does not have to be all or nothing, let us present a more specific but, nonetheless, very smart system. It is Weka: a fantastic datamining suite to search your data for those essential patterns that are more important in your life. And no secrets inside!. It is all open and shared at sourceforge. ​

Our final suggestion to look at (out of an enormous space of search) is a new project to integrate general intelligence into a computer. It is about a new family of programming languages known as LPS, (from Logic Production Systems). LPS aims to provide for easier languages to program computers, usable on flexible, web-based platforms, for tasks that require human-like intelligence, such as simulations of the prisoner’s dilemma and its alternative, equilibrium solutions.

The project is developed by a group of logic programs, led by one of the founding fathers of AI, Bob Kowalski. And it is shared at Bitbucket with an invitation to join in. All these three projects are all open source and their software run on Linux and Ubuntu. How about that as a sharing exercise? FLOSAI is invoking the power of decentralized collaboration. Would not be one of the finest ironies of human history that the perfect slave resulted from the greatest exercise on freedom and collaboration?

Reference

McCarthy, John; Minsky, Marvin; Rochester, Nathan; Shannon, Claude (1955), A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence retrieved Sept, 16, 2016.