Symbiotic Intelligence: The Future Where Humans and AI Unite

Transforming Fiction into Function: The Technology That Will Change the World

In my last article, I took an honest look into AI, identifying its obvious dangers and benefits. The simple conclusion was that AI is not inherently good or bad; its impact depends on how it is utilized.

One topic that received a lot of feedback and questions was the idea of personal AI symbiosis — machines enhancing us individually. Let’s take a closer look at this seemingly utopian scenario!

The Vision of Human-Machine Symbiosis

Considering human-machine symbiosis, yes, it sounds utopian. What could it look like? We don’t have to be sci-fi writers to explore this topic; let’s instead approach it from a logical standpoint by breaking it down into its essential components!

We can reduce every machine symbiosis into a simple input-output model:

Perception (Input):

We get data from AI. It can give us filtered, moderated, and summarized information, helping us better understand our surroundings — our private and professional social environments, politics, health, science, etc. The machine would collect, analyze, simplify, and explain data. In other words, we could receive AI-curated information about everything. AI would preprocess data and present it in a conclusive, accessible, and understandable way.

Action (Output):

We want to do something? AI can do it for us. It could help us act in our best interest. It could support us in our jobs (to keep them, even if competing with others’ AIs) and help us in our personal lives, whether deciding what to do next weekend or making serious decisions about health, wealth, and our future. The machine would act based on our reaction to this data in a way that it thinks we want, trying to understand our intentions and act accordingly.

We, In Between:

And the cheese in this AI sandwich is us, personally. Embedded between perception and action, by machines. AI observes for us; AI acts for us. We watch and decide.

Science Fiction or Reality?

Sounds like science fiction? No, it is already happening when we interact with ChatGPT. Imagine you want to write a scientific paper; you might use it to do your research, such as finding and collecting your data. Then you could ask the AI to find statistically relevant correlations and add some literature quotes. That’s all on the perception side. When you are satisfied with the findings, you could use again ChatGPT to write the paper itself. That’s the action side. And in between, in this chat sandwich, are we.

How Could This Evolve in the Future?
(Steps On The Way To Symbiotic Intelligence)

Let’s explore the possible future symbiosis by looking again at the two components, perception and action, and how they could develop:

Perception Enhancement:

Our perception of the world could range from receiving summarized text information, hints, and advice to experiencing a fully virtual, sensory-rich environment. Maybe we will wear glasses (like Apple’s Vision Pro) and headphones that enrich the real world, or perhaps we will sit somewhere isolated, being in a completely virtual world. Maybe we will be connected by a brain-computer interface (BCs, like Musk’s Neuralink) and won’t even think about the fact that what we see, hear, and feel is actually mediated or even generated by AI.

Action Enhancement:

The actions, on the other hand, could be “translated” or improved actions we set or intend to set. And even without an exact intention, AI could independently perform actions based on our set goals. Imagine if the AI would automatically congratulate acquaintances on their birthdays, pay bills, or handle our taxes without even bothering us. Or if AI could automatically switch on the heating in our apartment while we are on our way home, without even informing us, without us even setting such a rule.

And what about personal life? Maybe one day AI will negotiate a romantic date for us with another person’s AI, where we just have to show up. Or discuss with all family members’ AIs to find a solution for a problem so compelling that we all accept it.

The Future of Symbiotic Intelligence User Interfaces (The I without the A)

User interfaces is where we will probably see the biggest changes in the next few years, leading up to symbiotic intelligence. Right now, AI interfaces are sadly in the state of 1980s GUI technology — chats. Literally, like old-school text adventures. With maybe some pictures in between, if we have the patience to wait for their generation.

This chat-like AI conversation modality will be probably replaced soon. I don’t think an AI-extended Siri++ will be the killer app. No, I believe we will head away from having AI as an artificial “person”, interacting with us. Future AI symbiosis doesn’t have to feel like an interaction; true symbiosis makes the two parts indistinguishably married together. Let’s stop calling it AI and instead refer to it as Symbiotic Intelligence.

Because AI without the “A” is the future. Yes, currently, the “artificiality” in a conversation with a computer is the novelty, much like how the “motor” was in “motorized carriages” over 100 years ago. Today, we simply call it a “car” and drive it. The future lies in intelligence that doesn’t feel artificial at all, just something that is always around us and quite intelligent. Therefore, the future UIs of such symbiotic intelligences will probably just be invisible, serving us with information without being perceivable, just picking up our intentions without querying or even annoying us with questions.

Symbiotic intelligence will probably be like perfect gloves — a layer between us and the world that feels so natural we don’t even recognize we’re wearing it. No artificiality anymore.

Symbiotic Intelligence — Micro-Transactions

By again differentiating between perception and action, we can drill down how symbiotic intelligence should interface with us. Symbiosis is not a discrete transaction like our interaction with a supermarket, where we go to, select, buy, and leave with the goods we need. Instead, it is a continuous interaction, existing together and performing micro-transactions constantly.

Current AI interfaces like ChatGPT and Siri resemble the supermarket model. We go to them when we need something, give a prompt, and receive a useful response. Symbiotic intelligence interfaces, however, will need to be much more micro-transactional, continuously reacting, adapting, and delivering.

Continuous Perception Enhancement

Symbiotic intelligence will therefore not present us with a cart full of perceptions or expect a detailed list of our needs. Instead, it will continuously enrich our perception in subtle ways, such as a slight glow to attract our attention to something important or a simplified view of something complex within our sight. This seamless integration means that the AI will be constantly working in the background, subtly enhancing our awareness without overwhelming us.

For instance, imagine walking through a crowded street. Symbiotic intelligence might highlight the safest path or draw attention to a friend approaching from a distance. In a professional setting, it could dynamically display key points during a presentation or summarize lengthy documents in real-time as you read them. And maybe even deliver the information via a biological interface as knowledge into our brains?

Continuous Action Enhancement

On the action side, symbiotic intelligence will act on our behalf in a manner similar to modern cars automatically shading parts of their headlights to avoid blinding oncoming traffic. These actions will be constant and adaptive, aligning with our intentions without requiring explicit commands.

The Flow of Micro-Transactions

Much has been said about gamification, and symbiotic intelligence will likely adopt this concept, but not by turning everything into a contest. Instead, it will facilitate a flow of micro-transactions, similar to games. When we play an open-world game, we interact constantly, and the user interface responds in a stream of adapted information. Similarly, perception and action in symbiotic intelligence will no longer be granular or discrete but continuous streams.

In an open-world game, the environment changes dynamically based on our actions, and we respond to these changes intuitively. Symbiotic intelligence will probably create a similar experience in real life. It will provide a fluid interaction where our perceptions and actions are continuously enhanced and adapted by the machine learning technology, leading to a seamless and intuitive coexistence.

Summarizing It

Symbiotic Intelligence therefore has to:

Reverse the concept of objectification or even personification of machines; instead, be invisible, a non-object.
Avoid discrete transactional interactions; instead, be in a constant flow of micro-transactions.
Enhance perception through subtle, context-aware modifications without disrupting the user’s sense of reality.
Anticipate user intentions, executing actions autonomously while ensuring alignment with user preferences and goals.

Excursus: It Doesn’t Have to Be Dystopian

Invisible symbiotic intelligence sounds scary? It kind of is. Because machine learning technology in such a setup would have plenty of opportunities to manipulate us. At the very least, it distances us from “reality” (whatever that is). It could easily trick us into perceiving the world as it wants and let us set actions as it prefers.

So, would AI be free to manipulate us then? Not really. Because the question would be: why should it? The training of AI is an optimization for a goal. And the goals large language models (LLMs) are currently trained for are to give answers we consider good answers. Note that it is not about the “truth” or the “right thing to do” or “the best for humanity” or even “the best for us personally.” It just tells us what we (or the majority of us) want to hear. Future multimodal models will probably have a similar goal they are optimized for.

The manipulation would probably come from outside. Others might want to trick our AI to present us with information and trigger actions they prefer. Sounds scary, but it isn’t new at all. Because a similar thing happens to us every day thousands of times. We call it marketing, advertising, etc. Someone wants to trick us into perceiving things the way they want and lead us to set actions they want us to take.

So nothing new here, just another player in between us and the marketing: the AI. Future marketing and other (perhaps even political manipulation) will therefore focus on manipulating our AI instead of us.

We might think, great, now we will have one day a superintelligent AI protecting us from marketing and manipulation. However, we should not forget that there will also be superintelligent AIs trying to influence our own.

Bonus: Symbiotic Intelligence, Principles For Product Design

Based on the concept of AI-human symbiosis, we (at ObviousFuture, developing an Intelligent Media Assistant) derived guiding principles for designing symbiotic intelligence-powered products that seamlessly integrate into our lives. These principles ensure that these products enhance user experience in a symbiotic way:

Invisibility and Seamlessness: Symbiotic intelligence should reverse the concept of objectifying or even personifying machines. Instead, these systems need to be invisible and non-obtrusive, seamlessly integrating into the user’s environment without drawing attention to themselves. This design principle ensures that the technology supports users without them consciously interacting with it.
Continuous Micro-Transactions: Unlike traditional AI systems that rely on discrete transactional interactions, symbiotic intelligence should operate in a constant flow of micro-transactions. This means continuously processing and responding to user needs in real-time, creating the most uninterrupted and fluid user experience possible.
Context-Aware Perception Enhancement: The systems should subtly enhance the user’s perception through context-aware modifications. By providing timely and relevant information without disrupting the user’s sense of reality, these enhancements should feel like a natural extension of the user’s own cognitive processes.
Autonomous Action Anticipation: Symbiotic intelligence should anticipate user intentions and execute actions autonomously as much as technically possible. This involves understanding the user’s goals and preferences, and aligning its actions accordingly to support those objectives without requiring as little explicit commands as possible.

Ultimately, the aim of symbiotic intelligence is to become an invisible yet invaluable part of our lives, enhancing our abilities and enriching our experiences without feeling intrusive or artificial.

Symbiotic Intelligence: The Future Where Humans and AI Unite

Transforming Fiction into Function: The Technology That Will Change the World