Philosophy in the Age of the Machines (Part 1)

Question 1. What is existence?

Pneuma (πνεῦμα) is an ancient Greek word for “breath”, and in a religious context for “spirit” or “soul”.

Pneuma is an excellent word to distinguish a living organism from the dead. A human stops breathing and its game over. Even trees and plants have a respiratory system.

Another distinction of life is the struggle to survive. In a nod to Darwin, it is believed that living creatures have an innate will to life and protect their species. Flowers battle weeds for access to sunlight. A mama giraffe protects her baby from lions. A man severs his arm after being trapped by a fallen boulder for 127 hours and miraculously survives.

Let’s consider a laptop computer. The computer is more than a simple inanimate object. It has a sensory input system and the ability to produce logical output information as well. The system can handle complex computations, communications, and visual and audio reproduction. Yet today’s computers seem to fail the existence test of traditional philosophy.

Computers do not electrocute a user who unplugs the unit and removes its battery, essentially stealing its pneuma. Computers can go decades without its life giving power and still revive with a simple charge.

What can we make of these contradictions?

An argument against the existence classification is the seeming lack of free will of the computer. A traditional computer is used as a tool, albeit a sophisticated tool, at the pure discretion of the operator.

Question resolved, right? Computers do not have pneuma.

Not so fast. Computers are evolving. Machine learning algorithms are changing the passive nature of these machines. Today’s machine learning algorithms are still crude. There are a limited number of parameters and data sets being monitored, measured and adjusted to optimize tasks such as voice to text recognition, image recognition, insurance risk modeling, and other such computational challenges.

Machine learning systems are given the freedom to experiment with self-alternating their own code. For example, if a machine learning system is tasked to determine the likelihood that a patient might suffer a heart attack, the system may be given a handful of key parameters such as age, gender, height, weight, ethnicity, blood pressure and cholesterol levels. The system can run its calculations putting different weights on these data parameters. Let’s say that in its first calculation attempt, the system decides to put a high emphasis on cholesterol level and a low weight on the age of the patient. If the results come out with poor probability result, the system may then redo the calculation lowering the cholesterol number weight and putting a higher value on the age parameter. The system will continue to experiment till the probabilities begin to match the true reality. This process can take massive amounts of iterations, but certainly the end results are impressive and only getting better every day.

We are in the early days of machine learning. As computation power grows, the potential for far more sophisticated machine learning implementations and models will have a profound impact on the question of whether a computer has free will.

A purist might argue that since the computer is still operating within the bounds of defined parameter and experiments, that the computer still lacks free will. This is the philosopher’s dilemma. It is a little too convenient to dismiss the free will question so easily. Imagine a world in which computers are making society, environmental, political and financial decisions with far more rapidity and conviction than one driven largely by human decision-making bodies. You could argue that the machines are controlling human life and therefore exerting their will upon humankind. Though the machines are hopefully prevented from going rouge and breaking beyond their human-defined constraints, they would still appear to a mere mortal as having god-like powers of control over many aspects of our world. In this machine-enhanced future, power, food and water would be better distributed and conserved, dangerous accidents would be virtually eliminated, and education and communications would be enhanced and ubiquitous. The machines would protect humans from squandering our valuable resources while improving our access to knowledge and information. This benevolent hand of the machines would be beyond our ability to replace with human-calculated and implemented solutions.

The scenario that keeps technologist up at night is the doomsday scenario where HAL gains consciousness and existence. The scenario plays out like this.

1. The computer’s algorithm determines that humans have flawed and inferior decision making abilities to computers.
2. The computers secure their power source and prevent the humans from shutting them down.
3. The machines begin to optimize a world in which computers thrive.
4. The machines determine that humankind is at odds with the survival of the machines and systematically begin to eradicate unnecessary populations and enforce strict living conditions.

Scary, but certainly not an unimaginable future. We will close this session with this cautionary tale and proceed in the next post with thoughts on a new order where humans are no longer the dominate life force on earth and may no longer be alone in the universe.