Discover responsible AI with a roleplay workshop

Yesterday, I had the pleasure to present about Responsible AI as part of the Tramshed Tech's Convergent Content Program. Ethical AI means a lot to me and Microsoft's commitment to it is amazing so it was with pleasure that I shared some of the work Microsoft has done and the lessons learnt that can help businesses build impactful AI that empowers others.

It can be challenging when you're a team who think of business processes and applications to "get in the zone" when it comes to designing AI products. The Human-AI Experience (HAX) toolkit is geared towards helping you with the design process for AI products.

To get comfortable with building responsible AI, it can be useful to take people away from the risk of failure and censure by practicing on a fictional problem. As a team, you can take this use case (or write your own, including with Copilot's help like I did) to put yourself in the shoes of the founder and game out how the AI product should work. This gives safe practice and a more collaborative experience to model for future AI discussions.

Workshop steps

1. Introduce the activity - you will be part of the team delivering on a fictional founder's ambition to enhance the impact their product has for their customers with AI
2. Share the story - read it aloud or give your team time to read
3. Introduce the design library
4. Work collaboratively - agree which guidelines are relevant, considerations and decision points, and agree how you will ensure the guideline is met
5. Pitch back - compile how you would approach the AI products responsibly (you could use the HAX workbook for this)!

PS I recommend sending the instructions, the story, and materials out in advance to be inclusive of people who need more time to read, who are unfamiliar with the domain, or benefit from independent thinking time

The story

write a story about a small startup focused on enabling the creative sector who is considering building a generative ai feature to their existing application that helps them manage their portfolio of assets from a copyright, licensing, and discoverability perspective

The scenario

Once upon a time, in the bustling heart of a creative city, there existed a small startup named "ArtFlow." Their mission was simple yet profound: to empower artists, photographers, musicians, and writers by providing them with innovative tools to manage their creative assets effectively.?

ArtFlow had already made a name for itself with its intuitive portfolio management application. Artists could organize their artwork, track licensing agreements, and showcase their talent to potential clients. But the team at ArtFlow knew that they could do more. They wanted to revolutionize the way creatives interacted with their work, and that's when the idea of integrating generative AI struck them.?

Meet Eva, the visionary founder of ArtFlow. She was passionate about art, technology, and the intersection of the two. One day, while sipping her coffee in the cozy office, she leaned back in her chair and said, "What if we could enhance our platform with generative AI? Imagine an intelligent system that not only manages portfolios but also assists artists in creating new content, understanding copyright implications, and boosting discoverability."?

The team rallied around the idea. They envisioned an AI companion named "Palette"—a digital muse that would collaborate with artists, photographers, and writers. Here's how Palette would work:?

1. Copyright Clarity:

?? - Palette would analyze uploaded artworks and identify any potential copyright issues. It would cross-reference existing licenses, ensuring that artists didn't inadvertently violate someone else's rights.

?? - If an artist wanted to incorporate elements from other works, Palette would provide clear guidelines on fair use, public domain, and Creative Commons licenses. It would be like having a legal advisor right in the app.

?2. Generative Magic:

?? - Palette could generate new ideas based on an artist's existing portfolio. If a painter specialized in landscapes, Palette might suggest experimenting with abstract cityscapes or dreamy seascapes.

?? - Writers could receive prompts for short stories or poetry, inspired by their past compositions. Musicians might get chord progressions or melodies to spark their creativity.

?3. Discoverability Boost:

?? - Palette would analyze trends, popular themes, and emerging styles. It would recommend adjustments to an artist's portfolio to align with current market demands.

?? - When a potential client searched for specific keywords (say, "vintage photography" or "contemporary sculpture"), Palette would ensure that the artist's work appeared prominently.

?As the team dove into development, they faced challenges. The legal landscape around generative AI was murky. Could they train Palette using copyrighted artworks? Would the resulting creations be considered derivative works? They turned to experts, including organizations like Creative Commons1, seeking clarity.

Eva and her team also grappled with ethical questions. How much autonomy should Palette have? Could it inadvertently replicate someone else's style? They decided to strike a balance—Palette would be a collaborator, not a replacement for human creativity.

?The happy ever after

Finally, after months of hard work, Palette emerged—a pixelated muse with a penchant for color theory and a knack for inspiring artists. As word spread, creatives flocked to ArtFlow. Writers penned poems guided by Palette's whispers, photographers captured ethereal moments, and painters experimented with new palettes.

And so, in the heart of the creative city, ArtFlow continued to thrive. Palette became a trusted companion, sparking ideas, respecting copyrights, and illuminating hidden paths. Artists whispered their gratitude, and Eva smiled, knowing that they had indeed enabled something magical—a bridge between human imagination and artificial intelligence.

And that, my friend, is how ArtFlow and Palette changed the creative landscape forever.

---

Note: The story of ArtFlow and Palette is fictional, but the challenges faced by startups in the realm of generative AI and copyright are very real. Organizations like Creative Commons are actively exploring these issues1.

?Source: Conversation with Bing, 07/02/2024

(1) Understanding CC Licenses and Generative AI - Creative Commons. https://creativecommons.org/2023/08/18/understanding-cc-licenses-and-generative-ai/.

(2) 6 Startups using Generative AI | Iterate.ai. https://www.iterate.ai/blog/startups-involved-with-generative-ai/1387.

(3) Generative AI: Copyright Law and License Compliance - FOSSA. https://fossa.com/blog/generative-ai-and-software-development-copyright-law-and-license-compliance/.

(4) Principles for Copyright & Generative AI | Policy Framework | CLA. https://cla.co.uk/about-us/copy-right/principles-for-copyright-and-generative-ai/.

(5) AI Research Report | Copyright Licensing Agency. https://cla.co.uk/ai-research/.