I recently watched Her, the powerful, humanistic story about a man who falls in love with his artificially intelligent operating system, and it inspired me to write down a few thoughts that’ve been stuck in my head about the emergence of autonomous experiences.

I think of an autonomous experience as the sum of physical, cognitive and emotional interactions people have with systems that can sense and act on their own (or, in AI lingo, systems that exhibit intelligent agency).

And specifically, I’m focused on what a user perceives and experiences as “intelligent” and “autonomous” rather than a philosophical or scientific qualification. After all, a user experience is owned by the user, not the system or its creators.

Back to the future

From this perspective, autonomous experiences aren’t exactly new. Clippy, Microsoft’s much-maligned Office Assistant, and the Roomba, iRobot’s widely-adored robotic vacuum cleaner, are two iconic products that introduced autonomous experiences into the mainstream over a decade ago.

But it seems like we’ve hit an inflection point. Over the past few years there’s been a wave of products whose value propositions hinge on the quality of the autonomous experiences they deliver. For example: personal assistants like Google Now and Tempo, contextually intelligent apps like Relate IQ and Emu, enchanted objects like the Anki DRIVE toy cars and the Nest thermostat, and of course good old fashioned robots like the Google self-driving car (oh, and several other robotics companies Google acquired).

And thanks to a bunch of enabling trends—cheap sensors and processors, open hardware platforms and communications networks, rich data firehoses and knowledge bases, advanced machine learning techniques and systems, accessible 3D printing and manufacturing—it’s becoming ever cheaper and easier to build products that can sense, think, learn and act in wide range of environments. It feels like we’ve only seen the tip of the iceberg.

Autonomy: a new experiential dimension

So why the emphasis on autonomy? Though it’s not the only property of artificially intelligent systems, it’s the one I expect to have the biggest impact on what we currently understand as “user experience.”

Autonomy breathes life into products. Where products used to purely react to our actions and behave consistently over time, they can now proactively engage us, act contextually and even learn new behaviors with experience.

This opens up possibilities for powerful new experiences, but not without introducing complexity for users. As humans we instinctively attribute agency to these kinds of products, and so we try to make sense of their state of mind as agents—the motivations, goals, thought processes, maybe even emotions that drive their behavior. As a result, the user experience is shaped not only by product interactions but also by a deeper understanding of the agent experience, which is reflected through questions that users may now ask like:

• Why does the product do what it does when it does it?
• How does it makes sense of the world around it, including the user?
• What it is doing when the user isn’t interacting with it?
• How does the it change (or not) with time and experience?

This interplay between user and agent is both the challenge and opportunity of designing with artificial intelligence. AI isn’t just behind-the-scenes technology, it’s deeply woven into the the fabric of the user experience.

Exciting new frontiers for design

Autonomous experiences open up profoundly important frontiers for designing and building products. They have the potential to transform our behaviors and relationships with the things and the people around us, and even ourselves. And they require their own brand of intense, multi-disciplinary collaboration amongst product teams. Wired sums it up nicely:

AI: the ultimate UX challenge

This isn’t entirely uncharted territory. The HCI and HRI (Human-Robot Interaction) communities have been exploring this domain for decades, motivating a large body of academic work and commercial efforts of all kinds. Now entrepreneurial and corporate interest (or maybe just Google single-handedly) will help accelerate experimentation and learning.

The future depicted in Her may be a little ways off but the path towards it is coming into view. I’m super-excited about the journey ahead.