On meaning, machines and the stories we use to understand the world
There are some events so large that eventually we come to know them as a single story.
March 15, 2019. The Christchurch mosque terrorist attacks.
Fifty one people murdered. Forty more shot and injured. Two mosques attacked during Friday prayers. A city placed into lockdown.
Like September 11, the Boxing Day tsunami or the earthquake and tsunami that struck Japan in 2011, the scale of the event eventually produces an overarching account. We learn what happened. We learn the numbers. We learn the sequence of events. We remember where we were when we heard.
But inside every enormous event are thousands of much smaller stories.
This is one of them.
About a year after the Christchurch attacks, I was working at the National Library of New Zealand. I was responsible for public engagement, which included our public programmes and the spaces people entered every day.
I travelled to Christchurch with some of my leadership colleagues and we met with peers who had been working in public institutions in the city on March 15.
I wanted to know what they had done.
When a city suddenly goes into lockdown, what happens inside a library?
People cannot leave. Nobody quite knows what is happening. Information is incomplete. People are frightened. Staff are frightened too.
What helps?
One of my colleagues gave me two pieces of advice.
The first was surprisingly practical. Always have a very large supply of lollies. During the lockdown, she said, they had gone around offering them to people. Of course, you don’t need a lolly to ask someone if they are okay. But somehow the lolly made the interaction easier. It gave people a reason to approach one another. It was something small to offer when there was very little anyone could actually do.
But it was the second thing she said that really stayed with me. Know who among your staff and community are people of faith.
I asked what she meant. Did the particular faith matter?
No, she said. Christian. Muslim. Jewish. Buddhist. Whatever.
What seemed to matter was that, in a moment of profound uncertainty, people gravitated towards people who had faith. I’ve wondered since what they were looking for. Meaning, perhaps. Reassurance. Hope. Calm. Someone who could sit alongside fear without becoming overwhelmed by it themselves. Someone who could listen with empathy, offer comfort, or simply be present when there was very little useful to say.
Perhaps there was something reassuring about being around someone who appeared to have a framework for understanding the world when, for everyone else, the world had suddenly stopped making sense.
People who are deeply grounded in a belief system can sometimes carry a particular kind of steadiness. A sense that the immediate moment, however terrible, is not the entirety of the story. That there are values to hold onto, rituals to fall back on, communities to gather around, and ways of being with suffering even when it cannot be fixed.
Perhaps that was what people were seeking. Not faith itself, necessarily.
But some of the things faith can provide: meaning, connection, hope, compassion, ritual, continuity and the possibility that this moment was not the end of the story.
The questions underneath the questions
I thought about that conversation again recently when I came across a call for fellows from the Center for Humane Technology. Its areas of inquiry included AI and cognition, relationships and surveillance. But one stopped me.
AI and spirituality.
It makes sense when you think about it. People are no longer using artificial intelligence simply to ask factual questions.
We ask: Who am I? What should I do? Why do I keep doing this? Does this relationship feel right? How do I know when it is time to leave? What makes a good life? Does any of this mean anything?
These are not particularly new questions. Human beings have been asking them for thousands of years.
What is new is who, or perhaps what, we are asking.
The first time I remember asking AI a question of that kind was a couple of years ago. A friend gave me a prompt instructing my AI to undertake a Johari Window analysis of me based on everything it knew.
It felt a little familiar. Somewhere between a personality test and an astrology reading. The answer was useful. I repeated the exercise for this essay.
Google gave us access to information. Social media gave us access to an audience. AI gives us something rather different. It replies. And, over time, it can begin to feel as though it knows us.
That distinction matters.
A search engine gives me a list of places to go. A conversational system receives my question, processes it in the context of what I have told it, and responds directly to me.
Then I reply. It replies again. And somewhere in that exchange, information begins to resemble relationship.
Is that necessarily a bad thing?
There is, understandably, considerable anxiety about this. Some of it is justified.
A system that feels intimate can invite extraordinary levels of disclosure. We can forget that disclosure to AI is still disclosure. Information given in one context can become sensitive in another. Systems may increasingly be able to infer things about us that we never consciously chose to reveal.
There are questions about dependency, manipulation, surveillance, commercial incentives and the extraordinary concentration of power in the companies building these systems.
There are also legitimate questions about what happens when someone who is distressed, vulnerable or unwell begins treating a machine as an authority.
We should take all of that seriously.
Humans are not neutral either. We bring ourselves into every conversation. Our histories. Our experiences. Our prejudices. Our values. Our understanding of what a good life looks like. Even with the best intentions, we interpret someone else’s experience through some combination of our own.
We misunderstand. We judge. We project.
This isn’t an argument for replacing human relationships with machines. Quite the opposite. The relationships I have with people I trust are irreplaceable.
I see a therapist and find enormous value in that relationship. I don’t think a machine replaces therapy, nor do I want it to.
But there have also been evenings when something has been turning over in my mind and I have opened my AI assistant. Sometimes I need to get a work problem out of my head or dissect a difficult exchange. Sometimes I want to test my thinking. Sometimes I simply want to put something into words and have something respond. I also like being able to capture a moment of inspiration for something I might write later.
There is immediacy in that interaction. And sometimes immediacy is what I need.
But not always.
There have been other moments in my life when I wanted an answer, a decision or a resolution immediately, only to be grateful later that circumstances forced me to wait.
Time does something too. It creates distance. Emotion settles. Perspective changes. Things that felt urgent become less so. Sometimes what looked like certainty turns out to have been a particularly intense feeling on a particularly difficult day.
Perhaps there is value in a little friction. In not always getting the answer immediately. And I think this is something we need to watch in our relationship with AI.
Immediacy is one of its great attractions. But what happens when there is no natural stopping point? When the question can always be answered, the idea immediately developed, the next version produced?
I notice this in my own work.
AI has increased my productive capacity enormously. I can move between multiple projects and ideas, research something quickly, test an argument, develop a piece of work and return to something else without losing nearly as much momentum. Work that once took days can sometimes take hours.
I have regained time. But regained time is very easy to fill. The machine can keep going. My brain cannot. For now, I am managing that reasonably well. But there are moments when I can feel the strain of holding more projects, more ideas, more possibilities and more decisions in motion at once.
We have concentrated heavily on what AI might do for productivity. Perhaps we need to pay equal attention to what increased productive capacity asks of human cognition.
And perhaps we will need to design some friction back in. Ways to pause, test, reconsider, sleep on something or simply leave it alone.
The ability to do more does not necessarily mean that doing more is good for us. Sometimes the pause was doing work too.
Whose intelligence?
There is another kind of context at work here. Over time, there is now a body of interaction sitting behind my conversations with AI.
The system has context. It can say, effectively: this is similar to something you have been wrestling with for a while.
That can be extraordinarily useful. But it raises another question. What else sits behind the answer? We know these systems can reproduce bias because they are built from human generated data, human choices and human institutions.
Their answers don’t arrive from nowhere. If increasingly powerful systems are going to mediate how people encounter knowledge, whose knowledge will be present?
I don’t think the answer is simply to withdraw. Participation doesn’t guarantee influence. But absence has consequences too.
If people and communities with different histories, knowledge systems and worldviews aren’t present in shaping this technology, what will be missing from the intelligence we eventually rely upon?
There is an old saying I use at the beginning of many of my presentations:
Seek advice from many, but act with your own wisdom.
Whatever its origins, I find myself returning to it increasingly. It feels almost perfectly designed for a world of AI mediated information. Because perhaps that is the thing we need to protect most. Not our ability to receive an answer. Our ability to judge it.
Who gets to make meaning?
This is perhaps where spirituality becomes particularly interesting.
When someone turns to a priest, kaumātua, imam, rabbi, therapist, friend or elder with an existential question, they are not entering a neutral space.
They are entering a tradition. A worldview. A body of knowledge. A set of values. A particular understanding of what it means to be human.
Indeed, it is often precisely why we seek those people out.
The person of faith in the Christchurch library was not valuable because they possessed some universally agreed answer to catastrophe.
Perhaps they were valuable because they had a framework within which catastrophe could be encountered. A language for grief. A practice for sitting with uncertainty. A belief that the terrible thing happening in front of us was not necessarily the totality of existence.
AI also answers from a context. The difference is that its context can be much harder to see. Its responses emerge from enormous bodies of human language, design decisions, safety systems, commercial choices and cultural assumptions.
So perhaps the important question is not: Can AI give us meaning?
It is: What meanings are being made available to us through AI, and who shaped them?
That feels like a much bigger question.
Crossing the Delaware
Years ago, I visited the Metropolitan Museum of Art in New York.
I was on a Museum Hack tour, a slightly irreverent alternative to the traditional museum tour. The company apparently began when its founder took a woman on a date to the Met and, wanting to impress her, showed her the works he loved and told her their stories.
We were standing in the American Wing, near Emanuel Leutze’s enormous painting Washington Crossing the Delaware. George Washington stands improbably upright in a boat, crossing the icy Delaware River during the American Revolution.
Our guide was telling us stories about the works around us. And I remember looking across the gallery. Nearby was an African American pastor with his own group.
He was looking at the same art. The same paintings. The same collection. But he was telling different stories. I found myself watching him.
Our guide was telling us one set of stories. Across the gallery, the pastor was telling another. Same room. Same collection. Different histories. Different experiences. Different ways of seeing.
I realised that none of us encounters an object without context. Perhaps we don’t encounter anything without context.
Maybe that is what those frightened people in Christchurch were searching for when they gathered around people of faith. A way of holding what had happened. A frame large enough to place the incomprehensible inside. We have always looked to other people, stories, faith, whakapapa, art and institutions to help us do that.
Now another kind of intelligence has entered the room.
Perhaps the question isn’t whether we should ask it what things mean.
Perhaps the more important question is whether we understand the context from which its answer comes.
