AI Design Summit: Part 1 (Megan)

This year we attended the virtual AI Design Summit Conference hosted by ADPList, where industry leaders in Design gathered to share their unique insights and how they leverage AI in their workflow processes. As AI continues to gain momentum, these thought provoking discussions are key to elevating the solutions we create as industry best practices continue to evolve.

Unleashing Creativity: Real-World AI Techniques from CRZY Design Studio

Speaker: Stephen Gates

Stephen Gates kicked off the conference by discussing AI's role in his own design workflow while sharing helpful resources to stay updated on emerging tools. One central idea that he stresses throughout his AI experience is the need for human assistance, and encourages designers to collaborate with AI rather than relying solely on it. He aligns this idea with his own six-step workflow when generating photos, while sharing some tools along the way in different steps of the process:

1. Creating images: Midjourney
2. Refining the output: Magnific.AI
3. Finishing touches: Gigapixel.AI and Adobe Firefly
4. Repeating the process with the final result: Midjourney
5. Animating with video: Runway and Luma
6. Evolving your resource toolkit: Toools.design 

A great expression he shares during his talk is that "great work comes from great prompts that have great structure."  Without the correct understanding or knowledge of how to define what you're looking to achieve, the disconnect can often do the opposite of what the tools are meant to do in the first place — enhance the work you produce while saving time and energy.

This begs the question, how can we work to create a good structure to get the right output? He references one often used in his own prompts to help direct the end photo as intended. This includes defining the medium (photo, painting, illustration), the subject (person, character, animal, city), the environment (indoors, outdoors, on the moon), the composition (portrait, headshot, wide angle), and style (movement, artist, time period).

However, there are also some downsides to be aware of and keep at the top of mind as experimentation with new AI tools goes on. Stephen mentions several of his own observations with photo creation, such as the often unwanted details or elements, the odd human hand representations, and the many iterations and time it can take to achieve your desired result. He also cautions about the ethical implications and considerations in AI Design, and how our role as designers plays an important part in ensuring ethical practices and eliminating predisposed biases. Stephen concluded his talk be urging designers to join new communities, upskill, and experiment with new tools to ensure ongoing growth in a quickly changing landscape.

AI Inside + Out: Enhancing Products and Process

Speakers: Mike Gadsby, Mark Angelow, Simone Dehel, and Christine Sheller

Several team members from O3 World gathered to reflect on their experience of how AI has impacted the way they work. They sparked many questions that are important to reflect on when evaluating whether to integrate AI into your own process. These tools are like pieces to the puzzle, so we must ask the right questions to find where it could fit in. Do the AI investments make sense? Where could it fit in to save time and effort, and where could it hinder the process altogether? Organizational needs differ from company to company, so assessing how it could benefit the specific work you do is essential. 

Some of the tools the team utilizes are ChatGPT, Claude, Formless, and other AI Agents. Many of these they recommended and found the most benefit for use throughout the ideation/discovery and prototyping stages, as it can assist in generating content for wireframes, pulling out insights from user interviews/data, testing prototypes, fostering concept development, and more.

The team also dove into four specific use cases where AI fits into their products: 

• Personalization: crafting hyper personalized experiences by understanding data input, spinning off the interaction process, and adjusting the output based on user behavior. 
• Conversation: utilized through Chatbots, which can understand the tone of voice or which visit it is for the user and shape how that conversation should look based on that information.
• Prediction: helpful to predict the next best action so the customer experience is timeless. This can be shown through product recommendations, encouraging users to continue their journey, or make other suggestions on next steps to take.
• Automation: beneficial for streamlining processes, reducing tedious tasks, and speeding up the output and quality. 

All of this to say, the end goal is always to elevate experiences and empower users. And one way to do that, as the team at O3 says, is to "leverage AI as your intern." When you approach AI in a way that recognizes the benefits and potential obstacles, it will help to ensure success when adapting through this new age. 

Thriving in the AI Era as a Product Designer

Speaker: Peter Zhao

When opening his talk, Peter provided an interesting thought: AI doesn't change what we do, it transforms who we need to be in the design world. Looking at it in a positive framework can help navigate uncertainty and view it as an asset that allows us to achieve things that weren't reachable before. The power is in our creativity, and the need for human guidance in AI continues to prove itself more than ever.

Some things Peter focuses on in his message and advice to designers is to:

• Embrace versatility: develop a holistic, adaptive, and critically aware design perspective.
• Focus on uniquely human skills: continually build on the skills of empathy, strategic thinking, collaboration, problem-solving, etc.
• Become AI-augmented: think about how we can collaborate with AI more efficiently by understanding what it is good at and where to adopt it in.

Along with this advice, he shared insights on certain areas he thinks will change and evolve. This could mean new ways we approach interfaces, functions, and layouts, and instead shifting to interfaces that are more dynamic, context-aware, and adaptive. A possible example of this could be interfaces built with more than buttons and elements, and rather reflect an interaction model that learns from our behavior and changes based on that.

Another topic that came up throughout the Q&A was relating to sensitive information breaches and privacy concerns with AI, and recommendations on how organizations should approach it. Many of the speakers throughout the whole conference warned against putting sensitive information into any open source AI tool (such as ChatGPT, for instance). Some organizations train their own small scale language models so they can use them for specific business needs they see fit, without having to jeopardize clients sensitive information and adhere to contractual agreements. Although AI is not perfect, diving deeper with these things in mind will not only help to see how you can improve current processes, but also create better ones in the future.