Microsoft AI
- Project
- Branding, Print, Illustration
- Creative Director, Designer
- Sam Schick
- Year
- 2019
Case Study
This was one of those projects, for us. Microsoft approached our Creative Director, Sam Schick, to lead Creative for their new Office of Responsible AI. Merely tasked with designing the Office’s launch, enablement strategy, and manuals for the Microsoft’s initial publication of their Responsible AI Standards – meant to review any use of artificial intelligence that might infringe on human rights, and align Microsoft’s business practices with its commitment to responsible AI and its core ethical principles – it was … a whirlwind experience.
Directing creative across many studios – from Bing to Human AI Interaction, and Microsoft Research – and working with leadership teams up and down the tree, and on the cutting edge of applied ethics in technology, we relished the task of giving shape to and systematizing something with such broad, critical power and public responsibility. A core strategic asset of Microsoft’s, this work represented a confluence of mission-oriented storytelling and actionable specifications — hardly a more exciting space to be working in right now.
The Goal
Senior Leadership, and the Office of the President, articulated a pretty clear set of goals: Design a specifications manual that is extensible, amendable, and the center point for a vital research loop. Create a manual that is cite-able, and could function as a toolkit, valuable in print but progressively enhanced as a digital product that would inform a compliance testing platform. * Design and develop a communications and identity system for the very practice of Responsible AI at Microsoft, one that lives respectfully within the Microsoft Brand while distinguishing the Office as a vital and unique project.
Responsible AI App
Prototyping a workable digital application for the Standards was the holy grail, the dream of dreams: We already hoped for a specifications manual that didn’t just specify best practices, but would be able to guide a champion (think of a case worker for responsible AI application) through a non-hypothetical, actionable responsibility lifecycle.
Our prototype provided gated access through the user’s Microsoft alias and clearance levels, and offered quick access to sensitive use reporting modules. New research – from applied datasheets, analyses, and addenda – is immediately reviewable, key for the various stakeholders using the app.
Above all, the Standard itself is made significantly more useful, navigable, and even highlightable and citeable. Referenced studies and data are offered in expandable footnotes that can be saved, shared, commented upon.
For a project so wrapped in security measures, discretion, and privacy, there’s much we wish we could share, but … can’t. What is shareable is just the tip of something very big, made of ice, floating in the sea (what is the word?), but it was the project of a lifetime, informed as much by earnest aspiration and immediate ramifications.