Balance in AI

In the competitive world of AI, machine learning has revolutionized what experiences technologists will plan and build for. How can we empower our workers to create products that are grounded in human needs?  My team and I created this concept video to inspire conversations around designing for AI. 

We address these talking points: How can we create experiences that make people (not the tech) be heroes? Show how AI can be contextually aware of users' needs, relationships and circumstances and respond accordingly? At Microsoft, our mission is to empower people to do more.  By starting with human needs — versus tech — we can create experiences that enhance people's abilities and relationships.  

Goal

Demonstrate how AI can serve people through intelligent and emotional signals. Focus on scenarios and guide a conversation through our video and on Microsoft's internal Designing AI site for employees.

Role

Created art direction, UX flows, motion prototypes for video editing. Collaborated on story script and narration with writers.     

Creative Process

The video's central theme is how we can augment a person's ability to be extraordinary.  It's the human, not the AI, that is central to our story.  In these stories, we highlighted AI scenarios with emotional nuances: How to express condolences to a colleague? Help a child? Interact and educate students?

Our concept video imagines AI tech in the near future. To make this world believable, I pitched an intro montage that includes people using apps that are powered by AI now and potentially in the near future (2-3 years out). This would help our viewers understand how today's AI, which is mainly focused on cognitive skills, could adapt and learn with more emotional capabilities that we show later on. The scenarios I proposed were based on real R&D tech to be authentic to our engineering audience. 

Intro montage: Conversational AI assistant notifies user of travel delay and offers an alternate flight to fly.

Intro montage: Dad uses a medical app to scan his child's ailment to get info and medical assistance.

Art Direction

I collaged mood boards to give our film agency a feel for casting and environment and to inform UI motion animations. For stories that required UI walk-throughs, I created screen flows that I then made into prototypes to share with the motion designer to riff from.  

Mood board: UI is contextual to users and their environment. 

Mood board: Show diverse characters in age and backgrounds for inclusiveness. 

For cinematic UI, the challenge to design is to make them glanceable and easy for viewers to understand for the few moments they are on screen. Can the UI tell a story and enhance the narrative? 

For our student storyline below, our character is researching Harriet Tubman, the Civil War abolitionist, for a school paper. We imagine a world where these historical figures could come to life as bots — using AI and machine learning capabilities — to engage users with stories of their lives. To make our Harriet Tubman bot feel relatable, I gave her avatar a large footprint on screen so the viewers (and character) can see her. This helps makes the dialogue feel more grounded in conversation between our character and avatar.  As the avatar speaks, I prototyped pulsating rings to indicate audio activity. Below her avatar, I collaged historical artifacts of Tubman's (books, maps, portrait) that rotate in view so that our viewers (and our character) can see a connection between Tubman's works and her speeches. I tweaked the prototype a few times, timing the animation with the script, before I shared it to our motion/video editor. Below, you can compare the animation in the film with the animation I prototyped — I'm happy with how that turned out!

Tubman_2.png

Impact

The video was showcased at internal Microsoft executive presentations and in the Microsoft Research AI conferences. It is also featured as a learning material on Microsoft's internal Designing for AI website.