AI has been on the horizon for decades. AI has been coming for a long while. Since the mid 1940’s when conversations on understanding the nature of intelligent thought laid the foundations for information processing, challenges and fantasy have been mixed with influence with ideas from many disciplines. It has always held the promise of imagined possibilities, infinite promise and defining what it means to be human. Philosophers from Gottfried Wilhelm Leibniz to Blaise Pascal, very early on reflected on the design of intelligent machines. Jules Verne, Isaac Asimov, Frank Baum (who wrote the Wizard of Oz) and many others imagined responsive devices capable of communication with human beings and supporting and challenging our deepest concerns on being human. (Buchanan, 2006).
Nobert Wiener’s work on cybernetics, W. Ross Ashby, Warren McCulloch and Walter Pitt’s work on neural networks, contributions from communication theory, mathematics and statistics, logic and philosophy, linguistics and of course John Von Neumann and Oskar Morgenstern in game theory have left their mark on AI and ML (Machine Learning). Both these fields have grown well beyond any of the individual contributors. The landmark paper in Mind, 1950 which lead to the landmark imitation game, known as Turing’s Test was a major turning point in this evolutionary journey. The name AI was actually given at the 1956 Dartmouth conference on Artificial Intelligence. However, as this discussion unfolds over the decades, when AI is described, at the core of intelligence is always the concept of continued learning.
Most of AI today, like it was in the 1960’s is based on semantic information processing. Language, understanding and translation were always thought to be the cornerstone of AI because of the computer’s ability to store and retrieve huge amounts of verbal data, phrases and massive dictionaries. Gradually, understanding began to creep into the landscape and language understanding and translations have moved AI closer to providing humans with non-human conversant assistants. Knowledge based systems have overtaken logic based paradigms. Slowly since the 1960’s MIT, IBM, CMU, Stanford and many other think tanks have helped move AI forward. Today AAI, the American Association for Artificial Intelligence, is a thriving association serving the AI community.
The term AI and intelligent human behavior is not easily defined. Generally, AI describes the process of machines doing work that would require human intelligence. The term generally includes investigating intelligence, problems solving and creating computers systems that are intelligent. Sometimes AI is described as weak or strong. Weak AI implies a computer is merely mimicking cognitive processes and simulating intelligence. Strong AI implies computers are self-learning and intelligent. Computers can understand and optimize their own behaviors based on prior knowledge or data and experiences (Gerlind Wisskirchen, 2017). Other ways of describing AI is narrow, broad and channel. Narrow AI is the ability of AI to handle one specific task for the purpose of duplicating or replacing human intelligence. Diagnosing in radiology or skin cancer is often used as an example of narrow AI. Broad AI is systems that are capable of exhibiting intelligent behaviors across many processes or tasks. Broad AI systems may even exhibit other aspects of human intelligence someday. Channel AI is even broader, more influencing and more expansive (Growth Stage Podcast, 2018).
