Originally Published July 3, 2024
“The honor of asking the first question is yours, Dwar Reyn.”
“Thank you,” said Dwar Reyn. “It shall be a question which no single cybernetics machine has been able to answer.”
He turned to face the machine. “Is there a God?”
The mighty voice answered without hesitation, without the clicking of a single relay.
“Yes, now there is a God.” – Frederic Brown, The Answer
The industrial era taught us to be subservient to the machine. The digital era has held out the promise of liberating us from it. Instead of spending our lives in a factory, which is a machine building someone else’s creation, we can now create for ourselves. Yet, old habits die hard, and we insist on being in thrall to the machine.
This attitude has consequences. Instead of looking at AI as a liberating, creative technology, we envision it as a new form of master. Alongside attempts to use it to control humans more tightly, we get apocalyptic visions of future robot overlords. We moan about the wholesale replacement of creative industries and have developed an entirely new language of contrition: prompt engineering.
These fears, while not completely irrational, miss the point. The real breakthrough with AI is that now the machines are struggling to understand us rather than the other way around. Our systems and language need to adapt to that new reality if we really want to leverage the power over creativity that this technology gives us.
This is not a new thing. I wrote about this in Discovering Digital Humanity over three years ago before consumer-oriented large language models even appeared on the scene. It’s in keeping with an industrial mindset that humans must be controlled and organized through the use of machines.
This was true when relatively few large corporations or governments controlled most of the jobs. However, recent events have shown us just how much power we now have over our technological existences.
The pandemic laid bare the fiction of having to go into a physical location to get work done. Offices are a legacy of the shop floor. If anything, they are even more tightly controlled because “productivity” is even harder to measure when it’s paper, not automobiles. Even for knowledge workers, our offices were every bit as much the shop floor as Henry Ford’s Willow Fork plant was.
And how do you measure “productivity” in this system? The number of memos you generate is a poor reflection of your work output. Most productivity gains come from our technology and this was what allowed us to work from home when the pandemic forced the issue.
Hours spent in traffic getting to and from an office are also lost productivity. It’s just on the ledger of the employee and not the company, so it doesn’t count.
Lately, we have seen a lot of concern that businesses will think of AI as a way to control workers. This could make it into a truly oppressive tool designed to keep workers in place and following their every activity.
I think this is a minor concern because this is no longer sustainable. Like many of our systems, the 9 to 5 job never made sense (and neither did the class schedule).
The daily grind is a legacy of the assembly line. Production didn’t work unless teams of people were there to do their penance to the factory machine.
We have already seen technology breaking this system. People are quitting those companies and/or starting their own companies with powerful tools that allow them to compete with outdated business thinking.
This has been going on for decades. We’ve seen giant print houses undermined by desktop publishers who operate out of their living room. I don’t need a massive publishing contract to publish books anymore. I can have a gallery show of my photography on Flickr. There are relatively few businesses that require size to be effective these days.
The one area where large knowledge businesses such as the big consulting houses have an advantage is in their access to decision-making tools. These tools mainly work because their clients struggle to see the world for what it is. So far, at least, their product requires the sweat equity of teams of specialists.
As an entrepreneur, I cannot afford to hire McKinsey to build a strategic plan for me. However, I can use generative AI for the same thing. This is only the beginning, though. As I’ve highlighted in past blogs, AI has the potential to augment our ability to see and act in a complex world.
The next steps are to humanize tools, not to mechanize humans. Training humans to talk to the machine is forcing humans to become machines.
AI can be a profoundly humanizing technology because it requires a far lower level of technical skill than existing digital tools. We will unlock AI’s true power if we focus on creating interfaces and humanized communication methods with the models. By humanizing the technology, we augment ourselves, not systems or machines.
Conceptual AI-Driven Concept Map
My IdeaSpaces project is focused on humanizing our interaction with information at scale. Humans see in pictures. Maps go back before the invention of writing. They are profoundly human constructions but require a great deal of technical skill to make and remake. Up to now, machines have struggled with that.
This is where generative AI comes into play because it can generate and regenerate connections at a high rate of speed. Right now, what is lacking is the interface that turns those connections into maps of information.
This kind of tool will put great power in the hands of the user. It will turn consultants into teachers. The job here is to show the world new things and new ways, not to force technology and people into models that are inhumane, even if they were efficient for a time.
Systems are conservative, and it is a natural reaction for them to entrench in the face of uncertainty and change. We see this in people being forced to go back to offices. We see this in the way business and education systems have adopted generative large language models.
Everyone is doubling down on the familiar. It takes an act of courage to see the world differently, but that’s exactly what we need now.
Technology is inherently disruptive. You can’t get rid of technology or put the genie back in the bottle. It just is. How we react to technology is where opportunity lies.
It’s hard to imagine a world without personal computing devices these days, but imagine if everyone had just accepted the status quo in 1980 and rejected or regulated the use of these new devices by ordinary people. Some countries did that, but I rarely use North Korea or the Soviet Union as models for innovation.
We are at a similar moment. Our AI systems are more fragile because they lack a physical existence like an Apple ][ had. Any combination of reactionary forces can shut them down.
This is why distributed small scale open source AI is so important. Systems of humans will try to enforce the status quo in the face of overwhelming resistance. Centralized technology models like the most widely used large language models used today reinforce this.
The cult of the machine gives power to the few and they will act to preserve their control over the systems that put them where they are. Factories are examples of centralized power. They gave immense power to those who, as Marx put it, “controlled the means of production.”
We have replaced Big Steel and Big Oil with Big Tech. They all have one thing in common: they dehumanize our existences. This was not inevitable, nor is it necessarily permanent.
I prefer to think of technology as a liberating force for humanity. It has liberated us from the office and the factory floor. It can liberate learning from school. We just need to design our technology and our systems with human empowerment in mind.
If, however, we design technological systems to perpetuate our servitude, we will continue to be forced to pray to them. It’s time to declare our independence and build a more human future.