AI: adjusting the aperture
AI is a kind of telescope that opens the aperture of the human mind.
Let’s work on this for a moment in three distinct phases. Two we’ve surpassed, and one whose doorstep we are now upon.
First contact with AI came silently in the form of recommendation algorithms. “Collaborative filtering” is the technical term for how Youtube, TikTok, and other social or entertainment platforms suggest new content to users based on a constellation of connections to like-minded ideas. If you watch cat videos, the Venn diagram this media lives within includes dog videos, so you may want to see these in your feed. In this way, first contact with an algorithm powered by machine learning “opened the aperture of the human mind.”
Second contact was jarring, public, and hyped in the form of chatbots. Built upon Large Language Models fed data to approximate a human mind’s neural networks, these algorithms responded to our queries with the innovative and the banal. Those who have used ChatGPT will say most of its output is unoriginal, decontextualized, and trite. Most. These same users will frequently admit that there remains a fraction of generated content that presents itself as novel and exciting. It’s useful and progressive. It contradicts our assumptions that chatbots can’t surprise us with their products. It proves to open our minds to new ways of putting ideas together and creating meaning from our language.
I’ll pause for a moment to reflect before reaching our impending third phase. First contact was externally imposed control upon the user’s aperture. Youtube’s recommendation algorithm had an agenda (brilliantly unpacked in this difficult, malaise, and true narrative audioseries, Rabbit Hole) the viewer could not circumvent. This phase drove us to the extremes of our attention, awareness, and ability to discern truth. It continues (admittedly, with decreasing power) to make us consumers, haters, and fearful.
Second contact saw the user gain control of the “telescope” to use for their ends. The internet exploded with unique use cases, prompt generation techniques, and system-breaking queries. Arguably, the driver for this innovation was social media, an externally imposed purveyor of engaging content, but the fact remains that the power of creation was finally in the hand’s of the people. Whether this will turn out for good or ill leads us to the third phase of AI’s development in the West…
Third contact will see the exposure of LLMs to the whole of the internet, society, and the natural world. In April, integrations were released to ChatGPT which allowed the program to execute on other websites and query user’s chosen databases. Think of everything the “telescope” showed us before as two-dimensional, with this shift opening our eyes to the third dimension. The evolution of this technology heralds massive shifts for the way we accomplish tasks in our work, interact with company’s services, and understand our environment. It brings to light ideas and processes we never thought possible. It fully vests power in the fingertips of millions to better create meaning out of the world available to them online.
For all of the danger I write about this technology, we must keep in mind the positives we should seek to foster with it. Contact with a new mode of thought has always transformed humanity for the better. Language allows us to understand each other. Writing allowed us to understand time and pass down our histories. The internet allows us to understand the scale of our species and collaborate in ways that are impossible in a print world.
Many have called AI a mirror to reflect our own image back at us. Our hope should be that AI allows us to understand where our sense-making failed and how we have lost the faith in the true narratives that shaped civilization to bring about human flourishing. Without this self-awareness, our ability to see the path to restoration will be blurred.