Possibilities

As humans we are unique in our adaptive flexibility. A big part of this flexibility is our imagination, our ability to envision alternate scenarios and consider multiple possibilities. We become aware very early in development of causes and effects. We observe our surroundings and as we act on our environment, we make note of the effects of our actions. When effects are not what we want, we learn and adapt our behavior to modify the outcome. This feedback loop generates a causal model in our brain of how the world works. The more we learn, the more we can modify our causal model and adapt our actions to create the outcomes we desire.

As scientists and researchers create Artificial Intelligence capabilities in computers, the challenges of replicating the human ability to converse and answer causal questions such as why something happens have led to insights about our own cognitive processes. They have found that simply relying on massive amounts of data and the ability to rapidly calculate probabilities is not enough to simulate human intelligence. Certainly, these abilities have contributed much utility, but another leap in understanding is required. These concepts are explained nicely in “The Book of Why” by Judea Pearl and Dana Mackenzie.

The history of science and statistical methodology informs the discussion and in some respects slowed the development of AI systems because of the choice to disregard causation and instead focus on correlation between variables as a more objective measurement. Because inference of causality requires a subjective causal model, scientists and statisticians intentionally avoided this path. Clearly, much has been achieved with the scientific method. We are aware that bias does creep into objective science as well, but efforts to eliminate bias are part of any well done experiment or research study. Hypothesis testing and statistical analysis of chance has led to great advances in our understanding in countless fields of inquiry. Yet, these methods limit the creation of true human-like intelligence in machines. It is understanding the ladder of causation and trying to recreate it in machines that is leading to the most promising human-like AI applications. The importance of this insight leads to an appreciation of our unique ability to consider “counterfactuals” or other possibilities. We are able to imagine alternative outcomes based on different behavioral actions or decisions we can make or did make in the past. You might otherwise know this as “second-guessing.”

As we come to understand causal relationships and try to modify our behavior accordingly to achieve desired results, we sometimes struggle to make the necessary adjustments. There are always obstacles and barriers to hinder behavior change. Pleasure, pain, effort, discipline, emotional or physical fatigue and time limitation are some of the potential hindrances. Sometimes our theories of cause and effect are not correct. We can make mistaken assumptions about the results of our action and not get the outcome we expect. This is an iterative, learning process. Evaluating our own causal inferences is worthwhile to make our decisions more effective. Errors of causal modeling are important to bad outcomes. This has particular application to Artificial Intelligence as well. As causal assumptions are built into computer decision algorithms, errors in these causal models can result in undesirable outputs. This is where care and caution are in order to find safe, effective and consensus driven models to guide computer assisted decision support. In our own lives, examining our assumptions about the nature of reality and the effect of our decisions and actions is part of effective self-determination, autonomy and goal achievement. Adapting and adjusting our approach when we make mistakes and imagining alternative possibilities is a gift of human intelligence.

Take advantage of your unique abilities as much as you can. Do this by keeping an open mind, be willing to challenge and modify your assumptions, learn continuously and expand your knowledge. Keep imagining the possibilities. The more effective you become, the more real they become for you. Love and accept who and where you are now, but keep fine tuning your instrument: mind, body, spirit.