The Rise of Perspective Agents
Decoding AI's Disruption on Sensemaking
“Any sufficiently advanced technology is indistinguishable from magic.”
— Arthur C. Clarke, Science Fiction Writer
On a scorching summer day in July 2022, some colleagues and I met in an air-conditioned conference room in New York. The meeting had a vibe of curiosity and expectation. Our team gathered to test a new AI-powered platform that was said to be like magic.
Annika McTamaney, a computer science major from the University of Colorado at the time, took on the role of “DJ.” McTamaney executed prompts from a team of creative technologists, strategists, and producers. She utilized GLIDE—Guided Language-to-Image Diffusion for Generation and Editing—a precursor to Dall-E, the groundbreaking GPT-3 text-to-image model in a whole new way.
Within the code, we first input a peculiar trio of words: Cheese. Bus. Moon. Almost magically, a photorealistic image materialized in under a minute. We saw a bus, its wheels crafted from cheese, cruising city streets. McTamaney kept generating more visuals. Some, like the cheese-wheeled bus, bordered on the absurd. Others adhered more closely to reality, echoing familiar ketchup bottles, brand logos, and depictions of people. We played like newbies in a collaborative art class for over two hours.
More than the images, my mind raced to the implications of what just happened. Together, we had a before-and-after experience. We left a world of media creation we knew for a new one with weird, mind-bending implications.
Beyond new means to create, I wondered about further diffusion and democratization of influence. These models allow anyone to bring new visions to life. As we saw with Web 2.0, more empowered media producers lead to a diffusion of influence. With generative AI, who we follow or what we turn to for insight or guidance carries new questions. If influence becomes further "atomized" through generative AI use, what new forms of media and authority take shape?
Six months later, Open AI's ChatGPT and Dall-E gave us a glimpse into this disruptive potential. Hundreds of millions of users prompted new generative AIs to create digital images, text, and code with only words. It was the start of a monumental shift. Six months ago, generative wasn't a part of the cultural vernacular. Now, in the frenzied hype around the promise and peril of AI, generative engines like ChatGPT are an obsession and talking point in presidential debates.
A NEW ERA SHAPED BY PERSPECTIVE AGENTS
In recent years, what began as our team prompting cheese buses in GLIDE evolved into a futures lab integral to our business. Our core objective is to explore the impacts of emerging technologies, such as generative AI. We collaborate with leading universities, digital think tanks, venture capitalists, and AI studios to sense how bleeding-edge innovators will mold the present-future social landscape.
This landscape is fast-evolving and fragmenting. Ethnographic research we conducted with the Institute for the Future back in 2020 found accelerated, individualized action to sustain relationships, rituals, and a semblance of order. The public en masse used new technologies, platforms, and avatars to adapt to a jarring global pandemic.
The research uncovered overwhelming evidence of a significant cultural and behavioral shift. Through algorithmically primed social networks, QAnon attracted millions to build depictions of reality collaboratively. Witchtok sparked a vibrant witchcraft community on TikTok, with videos watched over 20 billion times. Trusted resources for COVID information came from automated dashboards. An app called Replika, billed as an AI companion that cares, became a "digital friend" to over 10 million people. A virtual wedding in China attracted 100,000 remote well-wishers on the video-streaming site Bilibili. The list of strange findings could go on.
Beyond individual occurrences, our research unveiled a pattern of autonomous behavior. Tens of millions of small acts culminated in a seismic cultural shift. Present insights from our team suggest a persistent growth in both individual and collective autonomy. The emergence of new AI-driven agents and platforms will only amplify this trajectory.
We might not be emotionally and systemically equipped to handle new, transformative effects AI will add to the autonomy equation. In “The Coming Wave,” Mustafa Suleyman, the former head of applied AI at DeepMind, presents an exponential leap expected in the upcoming decade. He highlights that AI models have expanded at a rate ten times larger every year for the past ten years. Projecting forward, he asserts, “We're on a trajectory over the next five years to increase by 10X every year going forward, and that's very, very predictable and very likely to happen."
Interpreting this profound transformation driven by human and machine autonomy is the central theme of my upcoming book, “Perspective Agents,” set to release in January 2024. The book will be another important guide for those looking to make sense of — and participate in shaping — a new social order.
During my research, I saw AI-powered thinkers and platforms as a novel category of agents—perspective agents—that are poised to profoundly reshape our perceptions and behaviors.
Perspective agents are sources to make sense of a changing world. With the ascent of AI, these sources will take on a new complexion and weight. They encompass human experts, AI integrations, virtual environments, and interconnected collectives. Among these, the “super nodes” stand out as pivotal hubs, influencing our perceptions, beliefs, and actions.
Perspective agents range from individual players to collective bodies crafting a new narrative of transformation. Figures like Elon Musk, Mark Zuckerberg, and Sam Altman epitomize technocratic agents, shaping the contours of reality. They masterfully merge collective human intellect with the precision of machine algorithms.
These manifest in new knowledge agents like Perplexity and Reader and updated platforms like Google's new chat-enhanced search. Unlike the static pages of print magazines or the scheduled TV programs, TikTok dictates the ebbs and flows of cultural trends through its bite-sized, visual loops. Such AI-driven platforms introduce status symbols without beating hearts or real-life bodies. Computer-generated influencers like Lil Miquela hold sway over millions of followers, a stark testament to the changing landscape of influence and celebrity.
Beyond cultural sway, other pioneering agents harness open-source intelligence (OSINT), a democratized blend of intelligence and news monitoring. For example, The Centre for Information Resilience assembles intelligence into a public-facing Russia-Ukraine Monitor.
The Russia-Ukraine Monitor offers visualized intelligence, showcasing vibrant pins connected to authenticated videos, photos, or satellite images. Together, they detail conflict timings and locations, civilian effects, and military movements. Green markers indicate troop convoys; orange ones signal gunfire and explosive events; red pins emphasize civilian harm and infrastructure damage, while grey designates Russian firing sites and weaponry. Collectively, this map presents a world-changing event vastly different than traditional TV broadcasts or military briefings.
Ben Strick, director of investigations at the Centre for Information Resilience, said, “Originally, open-source was quite a sort of geeky thing to do. Now, we’re almost like the people’s army.”
SERIALIZING A BOOK ON SUBSTACK
Such agents, melding into a digest of information and interaction, represent a new frontier for perspective-building. It's a nexus of new creation, curation, and content dissemination. These agents reflect a reformulation of media and tools for thinking.
The breathtaking pace of change happening now can't wait on the production schedules of the publishing business. Before the Perspective Agents book release, I'll regularly post excerpts, game-changing agents, and perspectives from those featured in the book here on Substack.
The goal of sharing these perspectives is threefold: to offer a look into the tech forces shaping disruption, to foster discussions that elevate our capabilities alongside rapid technological advancements, and to provide readers with practical guidance for embracing the opportunities amid the impending turbulence.
The changes afoot are happening in real-time. Decoding them demands a sharp and discerning outlook and an enhancement of our sensemaking capacities. We'll need new sources, tools, and communities to help do so.
It's hard to think of a more important pursuit than altering our view to adapt to transformational developments on our doorstep.
My hope is that this newsletter and forthcoming book can be a guide to help you. Please subscribe and spread the word to others who might be interested.
Chris Perry, September 2023