“Technological systems contain messy, complex, problem-solving components. They are both socially constructed and society shaping.” – Thomas P. Hughes
“It’s the questions we can’t answer that teach us the most. They teach us how to think. If you give a man an answer, all he gains is a little fact. But give him a question and he’ll look for his own answers.” – Patrick Rothfuss
Every day there seems to be a new story about how AI is destroying education, work, or civilization itself. When I dig into them, they almost always describe a situation where AI is being misused or abused by human fallacies or systems.
It’s like damning hammers by highlighting every instance where someone uses one to beat someone to death instead of using them to build a home. The hammer is an apt metaphor because I often feel like I’m playing whack-a-mole when people bring these stories up to me to justify ignoring, banning, or fearing AI tools.
People are afraid and they have a right to be afraid. However, we must realize that those fears are of change and our own infirmities (we all have them).
This fear blinds many to the possibilities of any new technology, but the cascade of skeptical articles of AI is unprecedented in my experience of new technologies emerging on the scene. I am speculating, but I do think writers and journalists are particularly threatened by this technology because, for the first time, there is a tool that threatens the routine aspects of their jobs rather than someone else’s.
It is time to embark on a clear-headed exploration of how humans should approach this technology. AI can’t be banned. At this point, that would be like banning wind from an outdoor wedding.
Nor should we ignore AI. There is no disputing its likely impact and the potential for humans to abuse it. We also shouldn’t dismiss its potential to change the world for the good.
We need a human-centered framework for moving forward into a changed world. For this, I would suggest a set of mirrored principles to guide us forward.
1. |
AI should facilitate transparency. AI provides us with a tremendous tool for seeing information. This vision should be what we use it for, even to shine a spotlight on the various AI models themselves. |
AI fails when we use it to mimic transparency. It is not all-seeing and therefore we should never trust it to give us the full picture. Those who sell it to us in that way are selling answers that are often false and misleading, leading to much of the criticism of the tool as a whole. |
It is a lack of transparency that fuels criticisms of AI that center on it generating a torrent of misinformation. I constantly must remind people that humans already do this. The problem is not the information, it is finding a way back to its source. AI built on transparency will always connect information (see next point) and help us navigate toward finding our own answers, not give them (see #5 below).
2. |
AI should facilitate connection of all kinds. It should show us how information might be connected to other information. It should help us connect better with other humans in a chaotic, busy, and complex world. |
AI fails when we use it to substitute for human contact or trust it to connect information uncritically. If we use it this way, we risk isolation and blindness. Again, these are human choices that we make. Failure is a sign that we are failing to address deeper societal problems about education and emotional well-being. |
Technological isolation is not a new thing. AI may make it more appealing, but it is no different from those who use it to hide for whatever reason.
3. |
AI should generate surprise. I like “hallucinations.” When someone from another culture says or does something that surprises me, my first impulse is to try to understand why that is. AI should elicit the same response from us. This is how we learn. |
AI fails when we use it to dampen human differences and regulate normative behavior. AI is a natural pattern recognition tool. If we use it to turn the rich tapestry of human experience, knowledge, and emotion into an industrial gray carpet, we’ve turned it into a robot overseer. This robs us of humanity and our powers of inspiration and innovation. |
Critics of AI’s impact on labor often ignore the reality that the industrialized economy has dehumanized us for decades, if not centuries. This is not a new thing. Neither is automation replacing menial labor.
If we only teach humans to do mindless, repetitive tasks, they will be ripe for replacement by any technology. This is a societal choice and one that needs to be fixed in the social sphere.
4. |
AI should facilitate creativity. Humans are naturally creative and are capable of building amazing things but often lack the technical skill to be able realize them. AI often provides a ladder that allows us to tell our unique stories in ways that were impossible before. |
AI fails when we use it to replace creativity. Creativity requires an expression of the soul. AI is soulless, so if you want soulless art, it’s a good tool. I, however, would not call that art (and don’t even when humans create it). |
AI can mimic human creative expression but cannot replace it. Creative people such as myself embrace AI because it gives us another tool with which we can express ourselves. Most of the criticism I see in this area comes from creatives who aren’t doing very creative work in the first place and those who think they own our creativity. AI can never threaten authentic creativity.
5. |
AI must promote questioning. It has the potential to show us the world in ways that allow us to overcome our human blinders. |
AI fails when we rely on it for answers. Generative AI relies on associations at the root of its logic structure. Sometimes these associations can be definitive but quite often they are spurious. |
Learning requires questioning. AI’s power to shift perspectives gives us a powerful ladder to see through our conceptual blinders.
When you rely on anyone for answers instead of exploring for them yourself, you make yourself vulnerable to bias. I have seen many criticisms of AI dwell on this question of bias. This is only a problem if we accept its responses unquestioningly as answers.
No matter where we get information from, we must develop good skills to question information that is given to us even when dealing with pure human-engineered content.
AI is actually a powerful tool for examining bias at scale but only if we use it to question information rather than assuming what it is giving us is somehow magically unbiased.
As humans, we do the same thing but (hopefully) have the critical thinking skills to dismiss those associations that don’t make sense. AI lacks critical thinking skills. We should never forget that.
How we adapt and learn as humans and societies is far more likely to determine whether we thrive or wither under the pressure of technological change. These are the kinds of questions we should be asking ourselves, not tilting at the windmills of a technology that can’t be stuffed back in the bottle.
These five principles and their antitheses should guide us as we use these tools and, more importantly, adapt our human systems to their capabilities and challenges. At the end of the day, these are a product of human choices, not technological ones.