Google AI: Overview

For the majority of my time at Google I was Principal Designer in AI, working on features for consumer HW powered by on-device AI, and UX concepts for advanced personal AI agents.

Google AI: Overview

Original ID concept for Google Clips, 2016

Roles: UX Director [2016-2018], Principal Designer [2018-2022]

In my last few months at Google Creative Lab in NYC, I began working much more with colleagues in Google Research, focussing on the subject of “on-device AI”.

A group run by Blaise Agüera y Arcas had been researching how to get machine learning and other AI techniques to work on personal hardware, such as mobile phones - rather than running shared systems in the cloud where the phone / user became merely the client.

This On-Device Machine Learning allowed for fast and crucially private interactions with AI. I transferred over to Google Research as a UX Director to work on this fascinating area, with huge potential for HW devices.

Initially I worked on the innovative Google Clips camera [released 2017] - a wearable camera with an on-device AI that took photos and videos based on what it’s owner trained it to observe.

It featured in the V&A’s “The Future Starts Here” exhibit in 2018.

Google Clips On-Device AI Camera as released

During this period the team also worked on a number of concepts and features powered by on-device AI that made their way to market inside Google hardware, such as the “Now Playing” music detection feature in Pixel 2 [2017] onwards.

'Now Playing' music detection powered by on-device AI in Pixel 2

In 2018 I became a Principal Designer at Google AI, and began a project looking at potential personal AI near-futures called “Project Lyra”.

I directed this over 3 consecutive years, using a small budget to work with many of the UK & Europe’s most interesting design and prototyping studios including Superflux, Special Projects, Nord, Normally and Tellart amongst others.

The work, which centred on how to relate to personal AI agents continues to have impact within Google - as well as resulting in published patents.

Figure from a patent filing for an on-device AI interaction
Animation demonstrating an on-device AI interaction

In 2021, as Large Language Models and Generative AI started to move out of raw research and become feasible prototyping materials, I directed “Project Rheingold” - looking at early LLMs and generative AI as ‘tools for thought’ and collaborated on the early stages of Project Lamda, Google’s first foray into LLMs that lay the foundations for Bard and Gemini.

I was invited to give a talk at TU Delft in October 2022 where I was able to talk (in abstract) about the work I did during my time as Principal Designer, showing some examples…