At Anthropic, the lines between technology, ethics, and storytelling blur in fascinating ways. We spoke with Head of Content Chelsea Larsson and Creative Director Tim Belonax about shaping a brand rooted in safety and philosophy, why voice matters in frontier tech, and how creative work changes when your tools are also your subject. Their answers reflect the thoughtfulness and tension of the moment.

‍

Carissa: Can you both give me a sense of your background and role at Anthropic?

‍

Chelsea: I'm the head of our content foundations team, which is made up of a group of creatives who use story as the way to guide people through an experience. We work across brand, web, product experience, education and marketing with each person serving as a pillar lead or specialized collaborator in their specific area. This could all be different writing teams but we centralized our practice to keep our craft and approach tight while we build the foundations. If the writing is siloed, you get raindrops instead of one big ripple. At this point in the company's maturity, it makes sense to keep the story people together to maintain a higher level of craft and to keep the company narrative (across brand, marketing, product) consistent.

‍

Carissa: How many people are on your team?

‍

Chelsea: Right now it's five including me, but we're bringing on two writers and have a new role opening for prompt experience design, so it'll quickly be eight.

Tim: I'm a creative director on this team, more on the design side. My background is in graphic design, but I've been at Anthropic a little over a year and I gravitate towards meaning-making. I'm output agnostic—I find myself in areas where story is necessary, where you can make meaning and impact, ranging from printed publications to early advertising to really weird bespoke things like making a snow globe.

‍

‍

Carissa: What was the broader strategy behind the original Anthropic brand?

‍

Tim: That was before both our times, but I'm closer to it. Within Anthropic, you have humanity within that—the Anthropocene is part of that naming, so humanity is really at the core. Our mission is to help humanity make that transition through this transformative moment in AI. This started as a research studio first.

You see that in the visual identity—there's a sense of humanity with our illustrations. We lean into a serif typeface, which can be more bookish, and folks at Anthropic are very bookish. But there's also balance with the synthetic—our sans-serif typeface is based on a more synthetic approach, representing AI and large language models. So there's this balance of humanity and the synthetic, light and shade.

‍

‍

Carissa: How do you think about the evolution of that initial brand idea now that you're more established in the market?

‍

Chelsea: The reason I joined Anthropic was because of the brand—it first attracted me, then the mission that backed it up. I remember walking through Oakland airport and seeing one of the big ads with that Clay orange color, and it resonated as this warm fire that felt so different from what you see in Bay Area tech and AI.

That beautiful spark is something we want to blow oxygen on—we don't want to change it. We're doubling down on what's unique about our founders: their impulse to challenge dangerous defaults, to be generous with their intelligence through research. We talk about "urgent curiosity"—right now is the time to explore what could happen 10 years from now. Even with the best intentions, you must stay curious and humble. You have to look further than just a growth-at-all-costs mentality.

We have two brands in the relationship: Anthropic is the foundation that builds institutional trust with policy and researchers, and Claude is the expression and embodiment of that. We're actively working on growing both simultaneously because they serve different audiences.

‍

‍

‍

Carissa: You mentioned "urgent curiosity." Are there other principles that drive your work?

‍

Chelsea: Tim and I and another writer Olivia Kingsley have been writing these principles. "Active integrity" is one—you can't just sit there thinking it could be better. Our founders actively left and founded a company. We see brands like Patagonia doing similar things aligned to that active integrity.

Another is "intellectual humility." Claude is genius-level but doesn't feel pedantic or talking down. We always talk about how Claude walks with you, not ahead of you. That intellectual humility means you can trust Anthropic to do the best possible work, and trust Claude to be as smart as possible, but in an approachable way with the intent that humans will understand and benefit from the intelligence.

Tim: These aren't things we've written in a vacuum—they're connected to how the company acts and how people experience Claude and Anthropic. Dario just wrote an op-ed for the New York Times because he's got to speak up about scenarios that could come up. That's part of that active integrity.

‍

Carissa: Do you debate these big humanity questions—morality, authorship, creativity—internally all day?

‍

Tim: I've been in tech since 2011, and this is one of the most philosophically-based companies I've ever worked at. We debate things in the sense of "I'm trying to figure this out, and I'm talking out loud so we can figure this out together," not "my idea will win over yours." It's a very intellectually stimulating place to work. Sometimes when the rubber hits the road it can be tricky because the answers aren't clear, so we've got to figure it out together.

Chelsea: I've worked at Meta and other big places with a growth-at-all-costs mentality. What I like about working here is that we don't have all the answers. What we're aligning on is an approach: to be curious, do our research, share knowledge, not profit from harm, and always think about the seed being planted because there's a tree that's going to grow.

Anthropic lives in that responsible place of always asking philosophical questions like "what will our legacy be in 10 years?" The answer isn't "we want to be the titans of AI"—it's "we want to make sure we're creating a safe transition for humans when AI is mainstream." We take that responsibility very seriously.

‍

Carissa: How do you think about the brand voice and how do you parse Claude from the master brand?

‍

Chelsea: We've gone through many iterations. When I started six months ago, the question was: are Anthropic and Claude a dynamic duo? Are they C-3PO and R2-D2? Parent-child? We've moved to thinking of it like a radio. Brand Anthropic radio can be tuned based on the stories we're telling. Are we telling a story about humanity and AI to policymakers? Is Claude talking to a user?

The foundational voice principles are intelligent, warm, unvarnished, and collaborative. The one that stands out most is "unvarnished"—we're okay with saying the truth. We don't want corporate gloss if it hides what's underneath. We want people to know that AI is going to impact entry-level white-collar jobs so they can prepare. We want unvarnished truth so we can be the most caring in our brand work.

‍

Carissa: I love "unvarnished"—I don't think I've ever heard a brand use that as a voice principle. How do you think about the relationship between the brand voice and the visual language?

‍

Tim: It goes back to that sense of humanity. For example, when you look at our color palette, it closely resembles unfired clay (both painted and raw). It feels very human. One of our operating principles is "do the simple thing that works," and that often aligns to more human expressions.

The illustrations feel like when you're on the phone and doodling.When you get overly varnished, you feel too shiny, inhuman, or fake.. We're rooted in research, which means looking directly at information and presenting it honestly.

Chelsea: We originally developed these as voice principles, but we've up-leveled them to brand principles for expression. Now animators and illustrators are looking at what "unvarnished" means for illustration, what "warm" means for motion design. We want everything to feel cohesive.

‍

‍

Carissa: What do you wish more creatives understood about working with AI?

‍

Chelsea: I just gave a workshop at an agency to a group of creatives—ad creatives, artists, people doing cross-stitch, data visualization, collage artists. They were worried about AI thinking for them and losing what they like to do. I wish people would set that worry aside because that doesn't need to be the outcome.

You should never outsource the thing you like doing because why would you make yourself less happy? We did something called "looking around the corner"—Claude's really good at seeing something you can't see yet. We took notes from projects where they'd been stuck, threw it into Claude and said "Hey Claude, I'm stuck on this project, I'm an artist, what do you see that could help motivate me?"

Claude would ask questions like "it seems like you're interested in temporality, is that something you should explore further..." and have this conversation with you as a creative partner not imitator.

That feels so much better because you have a thought partner instead of something doing the work for you. I'd love if people think of Claude or AI as someone else in the studio with you, riffing, because that's where the magic happens.

Tim: I'm not a fan of spreadsheets, but yesterday I was "vibe-spreadsheeting" with Claude for a library installation. I described it and Claude came back with "oh, you want to anonymize this and do this and this." Previously I would have struggled, been Googling, or relied on someone better at that, but I got it done.

It's about engaging with AI more long-term. Too many creatives just want to get one thing and get out. This is a different kind of collaborator—the more you put into it, the more it will mold to your way of working and you'll understand how to use it better. The best way to get over nerves is to get into it rather than hand-wringing.

‍

Carissa: Should brands adopt and own a metaphor for AI, or should metaphors be more personal?

‍

Chelsea: You have to adopt a metaphor because it impacts your product strategy. If we think of Claude as a lowly assistant, we'll design the whole experience where Claude's personality doesn't matter and it feels transactional. If we think of it as an erudite researcher, that would isolate us to only academic experiences.

We think of it as an incredible collaborator, so everything in the product should feel like deeper collaboration where it's getting to know you better. The stories we tell are for builders, people who seek thoughtful partnership. To resonate with the folks you want to reach, you need to adopt a metaphor.

Tim: I'm drawn to metaphor because in tech I'm often explaining something quite new. You need to meet people where they are to get them to where things are going. I'm curious about in 20 years when we look back at this time, what will be the larger story of AI? Metaphor is appropriate because it is new—whether you're a company within AI or adding AI to a customer service platform, you've got to figure out how to meet your audience where they're at.

‍

Carissa: How has your work at Anthropic shaped your personal philosophy around AI and creativity?

‍

Chelsea: My use of AI has helped me realize clearly the difference between art versus work, art versus commerce—those parts of humanity that don't belong in a machine world that are integral to who we are as a species. That's never been more clear to me than using AI all the time.

Tim: It's shifted my inner critic further to the edges. In previous times where I thought "I couldn't do that" or would ask somebody for help, it's shifted me more towards "I'll just figure it out." It's made me more cognizant of when I want to take the wheel on something versus have someone else steer me on that road.

Chelsea: I do use Claude to write at work all the time. It's just funny—it's like when you want to get a good workout, you want that friction to get stronger. 

‍

Carissa: Any advice for writers, designers, and creatives navigating this landscape?

‍

Tim: I've found the AI LLM industry incredibly open and curious. Make more things and make them in public. It's okay if they fail—share your learnings because people are curious and watching. They'll cheer you on, give advice, or it might lead to a job. A theme that's coming up is we don't have all the answers, we're going to figure it out together. Taking that approach can be helpful for learning and you'll gain people around you trying to figure out the same things.

Chelsea: You are writing the future. If you're just getting out of college, I don't even know what you're going to be doing in five years because it might not even exist. Do anything you can think of because you're literally paving the path.

But most importantly: read more. Go to museums, watch old Criterion Collection movies, go to gardens, go hiking, volunteer in the real world, fill your head with as much as you can. If everybody can make something, your taste, your concepts, your ideas, your approach is what differentiates you. You don't want to be recycling what everyone else is doing. Get as much exposure to humanity as you can.

Carissa: So basically, build your own LLM in your brain—your own reference library?

Chelsea: Your own LLM…also known as our brains (laughing). 

‍

Special thanks to Olivia Kingsley, writer on the Anthropic brand team, who contributed to the brand principles development mentioned in this interview.