The Social UI
Technology change is so fast that it's nearly impossible to predict the future of anything anymore. Let's do the impossible and build a framework to anticipate where AI is headed.
We live in a world without a future. It’s not that we don’t have a future; it’s that technological change and the complexity, novelty, uncertainty, and disruption it has unleashed make it nearly impossible to envision the future with any clarity. As our time horizon shrinks from years to hours, it’s likely time to admit we’ve already passed the event horizon of a technological singularity (in physics, black holes are gravitational singularities where our understanding of physics breaks down).
Fortunately, if we focus on AI, a major driver of technological change, we may regain some clarity and reverse some of the confusion generated by this singularity. Of course, thinking about the future of AI isn’t easy either for the same reasons listed above. However, it may be possible to glimpse AI’s future by narrowing our scope to the user interface that a fully realized AI will use (the methods by which we interact with a technology). Here’s why a focus on user interfaces is important;
When a dominant user interface (UI) changes, big things happen.
A UI shift deepens and evolves our connection to a technology — expanding the number of users, the breadth of application, and the degree of use.
If we can build a framework for a user interface that will unlock the ability of AI to transform society, it should be possible to envision the future again.
The Network UI
Let’s start with a review of where we are right now. The UIs are network tools that leverage our interaction with networks (the Web and social networks).
Searching the Web to discover information.
Scrolling feeds for social networking involves finding new information (novelty) and optimizing pattern matching (sense-making).
In the past, search and scrolling were ‘automated’ by algorithms to sort, prioritize, and structure the information on the network.
AIs are something new. They are highly compressed models of the information available on the network. This attribute allows them to do nearly everything search and social networking automation currently does and so much more. It’s also why AI is rapidly eating the UI for search (Perplexity, etc.) and social (TikTok is an example of a highly evolved version of this).
Agents
However, AIs are far more than simply a replacement for current information automation. Their capabilities will allow them to penetrate every sector of the economy and society with new capabilities.
The first iteration will be AI agents, which will use text and voice — natural language processing (NLP) and language user interfaces (LUIs) — as their user interfaces.
These agents will provide a way to ‘automate’ a significant portion of the 80% of tasks in organizations and society that aren’t addressable by hand-coded software due to irregularities and batch size.
These agents will be trained as specialists and built to run on AI platforms that enable a basic understanding (language, voice, conversation), enabling individuals and organizations to accomplish categories of tasks that are not easily automated.
However, as advanced as these AI agents will be, they will still be ‘tools’ that offer a more malleable form of automation, not the transformational technology they will become. To become a transformational technology, AI requires a new UI.