SUMMARY
Elisabeth Carr offers four principles that help Meta teams strike this difficult balance.
Artificial intelligence (AI) has been around since the 1950s, but advancements in the last decade have introduced the most transformative implications for people. From the deep learning revolution to large language models (LLMs), AI has become more pervasive and accessible, changing how people engage with technology and with one another. AI has also played a critical role at Meta over the last ten years: It influences the algorithms that determine which posts you see at the top of your feed and protects you from bumping into furniture while using Quest 3.
As designers, we’re both the trailblazers and the tour guides, navigating the tremendous opportunity and responsibility that comes with designing for a future with AI. As the VP of GenAI Design and Creative, Meta, I’m thrilled to be at the forefront of this journey alongside my teammates.
Designers aspire to innovate with cutting-edge technology in a way that still feels approachable and accessible to people who aren't familiar with it. With AI, striking the right balance is even more of a challenge. In my role, I encourage teams to focus on finding the intersection of innovation and experiences that resonate and feel familiar.
It’s a tough balance to strike, and the stakes are often high.
Meta AI1, our AI assistant, has been and will continue to be millions of people’s first time interacting with AI. How do we help people connect more deeply with the things and people that matter to them using Meta AI, but also make sure our massive and diverse audience feels confident enough to use this new technology?
Use Meta AI assistant to get things done, create AI-generated images for free and get answers to any of your questions.
This is a design challenge we don’t take lightly. Designers at Meta can draw from extensive experience bringing new technologies to the world and the skills any designer should have in their toolkit, like creativity, craft, empathy and collaboration. Here are some of the principles our design team keeps top of mind to make sure our work in AI feels both future-facing and grounded.
Create a sense of familiarity
Today, when people talk about AI as a new technology, they’re often referring to generative AI (GenAI) that can create content like images and text from a combination of user prompts and large datasets. While it can feel novel and exciting, it can also be overwhelming and introduce unforeseen risks. It’s helpful to remind ourselves and other designers to set clear expectations of the presence and role of AI, and to do so in a familiar way.
One way we’ve tried to do this is by leveraging familiar messaging threads — a Whatsapp chat or Facebook message — but including clear signals of AI’s presence and limitations. We chat with our friends and family constantly and usually understand the nuances of their responses, including tone and perspective. When chatting with AI, however, responses may not always be relevant or accurate. Including purposeful and contextual messaging, as well as opportunities to give feedback, within existing interface-messaging gives people the same sense of familiarity without added risks.
Another way we worked to signal the presence of AI and build a sense of familiarity is through the Meta AI symbol. It has emerged across Meta as a designated signal for Meta AI capabilities. Whether using Facebook, Messenger, Instagram or Whatsapp, people tap the colorful icon to ask Meta AI questions, generate an image, or share an image to learn more about their surroundings. It’s the entry point for Meta AI across our apps, creating a sense of familiarity when you want to engage with AI or other tools within our platform.
The Meta AI symbol is consistently present across all Meta technologies that offer AI, serving as a familiar entry point.
Amplify human expression and creativity
At Meta, we're committed to harnessing AI's creative potential alongside its practical applications. We’re also ensuring designers and creatives are partnering with researchers, engineers and product managers on the development of our foundation models and product strategy. This collaboration allows us to innovate in ways that celebrate and amplify human expression and imagination. Our focus on creativity has led to the development of AI features that unlock new avenues for self-expression. One example is the ability to transform text prompts into vibrant images that reflect your imagination. People can generate unique, fantastical selfies in any reality they can envision — from a 1940s film noir character to a retro-futuristic space explorer.
People can use Meta AI to edit and transform images through instructional prompts.
People can further explore their creativity with AI Studio2 by creating custom AI characters tailored to their interests. Whether designing an executive coach for acing an interview or a movie buff to chat about Kurosawa films, these characters offer endless opportunities for fun, creativity and support.
Human imagination serves as a powerful catalyst for realizing AI's potential in fostering joy, play and self-expression. Our teams have the opportunity to influence creative tools and experiences for billions of people. As designers, we lean into our own creativity to lead this inherently creative work.
People can engage their fans or explore their interests further by creating custom AI characters through AI Studio.
Design for trust
GenAI models are capable of sophisticated reasoning and understanding. While this opens up new creative possibilities, it also presents concerns about loss of human control and agency. To avoid this imbalance, designers need to prioritize giving people agency and building trust with this emergent technology. When designing Meta AI, for example, user experience researchers (UXRs) spent months developing a framework for building trust with AI, identifying core trust-building behaviors to express in the AI assistant’s interaction with people.
Two of these trust-building behaviors are “listen actively” and “display care.” To listen actively, Meta AI will ask clarifying questions and play back information, helping people feel heard. Meta AI also acknowledges when it makes mistakes and asks for feedback, demonstrating a willingness to improve and take action. Making sure members of our community feel heard and reassured that we’re prioritizing their needs is foundational to building trust.
Designers, through a human-centered approach, naturally practice these trust-building behaviors, always considering human needs in addition to technical capabilities. We ensure this skill is leveraged when crafting AI features that are novel but also designed to feel familiar.
To build trust with audiences and improve the quality of its responses, Meta AI asks for feedback.
Stay open and collaborative along the way
I’ve never witnessed more cross-disciplinary collaboration at Meta than I have while working at the intersection of design and AI. UXRs, product designers, content designers, content engineers and sound designers each bring their unique skillsets to the same table to give people comprehensive, high-quality AI experiences. For example, rolling out Meta AI across our technologies has been one of the largest company-wide efforts in my 13-year tenure.
The company’s commitment to open source AI models and the recent release of Llama 3.2 reflects this spirit of collaboration. As our founder and CEO Mark Zuckerberg shared:
“With past Llama models, Meta developed them for ourselves and then released them, but didn’t focus much on building a broader ecosystem. We’re taking a different approach with this release. We’re building teams internally to enable as many developers and partners as possible to use Llama, and we’re actively building partnerships so that more companies in the ecosystem can offer unique functionality to their customers as well.”
We recognize that when more people can work on Llama, the models will improve. This collaboration that enabled open source will ultimately drive a more positive AI future for all of us.
Designers inherently understand and practice collaboration. We know that our work in AI has a stronger chance of resonating when we collaborate across functions, disciplines, organizations and even industries — and especially with those we’re designing for. If we want AI experiences to both innovate and resonate, opening the work up to be shaped and ultimately improved by others’ specialities and voices is key.
In this same spirit of openness and collaboration, I’m eager to share with designers working in AI and learn from others as we trailblaze this new frontier together. Anchoring ourselves in creativity, empathy and collaboration will always be vital to designing experiences that act as a bridge to the future. Fortunately, these are second-nature to designers, and we’re well-equipped to lead and light the way forward.
1Meta AI is available in select languages and countries only.
2Available in the US and in English only.
RELATED STORIES
Read more stories about process
Design at Meta is for everyone who touches user experience and design.
Whether you’re a product designer, writer, creative strategist, researcher, project manager, team leader or all-around systems-thinker, there’s something here for you.
