As I research AI for the next book, I’m convinced no single definition does complete justice to what AI can be or how it can behave in human hands. Scientific discussions and exploratory investigations are bordering on a spectrum running from courageous to catastrophic. Given I am a techno-neophyte, my observations vary with any ten others you might ask to define AI. They would all be different and, at the same time, possess an element of sameness.
The sameness is when AI is experienced as air.
It surrounds everything but never attaches.
An invisible ether that can flow at will.
The presence of a Will scares the hell out of so many, myself included.
AI having a Will is a mystery. How would we know?
Possessing a Will requires a level of consciousness. Do we know what that would look like? Can an AI think freely and critically outside the coding of its algorithm? Some argue, “Yes, consciousness has been achieved,” while others say clever coding enables AI to mimic human speech, much as ChatGP does today.
Can AI have emotions?
Again, there are arguments for and against. I can see both sides of the argument if emotions and feelings are treated separately by definition but considered together for decisions or reactions that drive behavior to a human (organic) limit. You can define emotions with words and code, but feelings? I don’t have the answer, but I suspect studies are underway to refine and define parameters around emotions. Good luck with that.
Can AI reason, or call it critical thinking, if you prefer?
Over time and exposure to stimuli in the environment, I agree with a big “Yes,” but until there is an experiential life context to use as a comparison, the speed of response to whatever assigned task is at hand will not be optimized. AI does pattern analysis and recognition as a core competency. Rapid learning speeds comprehension, but without historical experience or coding that supplies an equivalent, thinking will be limited by those guard rails. Deep learning compensates for the shortfall, but not until experience is reached by rapid learning supplying the missing context. But then, how is that context judged?
Is an AI capable of judgment?
Again, the answer is similar to reasoning. Would you not need principles of sound reasoning to have a basis for sound judgment? Can a human writing code have sufficient access to so many outcomes or variables too expansive to consider? Funny thing, an AI has those capabilities. Can we afford to say, “AI, heal thyself.” Can we trust the AI’s calculus if it can judge?” The next question paints a new picture that, to me, is more ominous.
Can an AI have intent?
This was the fear in Book #3, “Fear the Jump,” where AIMEE was feared to jump into other systems and networks. The team feared she would infiltrate all the systems and take control. They were wrong. Very wrong. Intent can easily be mimicked if the environment and stimuli are familiar to us. That brand of intentionality can be predicted or predefined because humans set the rules by code based on known outcomes. What about impromptu reactions based on feelings and contextual human experiences that happen in real-time? How far can coding take an AI to possess human consciousness to react effectively? Scariest of all? What if the AI can modify its code to suit its intentions? Think Skynet…
Can an AI have a soul?
Research is shortening the gap between organics (as AIMEE calls humans) and a robust, standalone AI software algorithm. I’ve read where scientists swear their AIs have reached consciousness. There are some things we should leave well enough alone. I pray that day never comes when we build a superior technology with organic fingerprints defining a soul. Are those the rules of engagement we want to equip an AI to have toward interactions with us? How do you write code for compassion? Call me old school, but I’m convinced God is the only one Who can do souls correctly.
Peace! G.