[0:00:20] We are back. It's not every day you get to hear from someone leading one of the most important conversations happening around the world. Not just about what AI can do, but what it should do. Daniel Amadei is the president and co-founder of Anthropic, the company behind Claude, one of the most advanced and fastest growing AI assistants in the world.
[0:00:47] With a background in public service, Daniela brings a grounded, human-centered approach to one of the most transformational technologies in our lifetime. Under her leadership, Anthropic has championed a new approach to AI safety known as constitutional AI, embedding principles like fairness, transparency, and accountability directly into the way the technology works.
[0:01:12] Recognized globally as one of the most forward-thinking leaders in the field, it's my absolute pleasure to welcome her to the stage to discuss the intersection of innovation, responsibility, and the future of AI. Please join me in welcoming Daniela Omere. Thank you. So, thank you so much for joining us today. Thanks for having me. It's great to be here.
[0:01:34] Claude has been growing like crazy. But for anyone that isn't familiar with Claude, can you explain what it is? Absolutely. So hi, everyone. I'm Daniela. I am president and co-founder of a company called Anthropic. And Claude is the name of the AI assistant and the models that we develop at my company. Thank you so much.
[0:01:56] So before we get into the big questions, I thought we'd start with something fun. A few rapid fire questions, if you're willing. Sounds great. Excellent. So just say the first thing that comes to mind. OK. Do you have any favorite or unexpected use cases of Claude? Yes. Claude was really helpful in party training my toddler. That is unexpected. Tell us a little more. Very unexpected. I don't know how many of you are parents, but I have a three-year-old and a negative four-month-old.
[0:02:25] Potty training is a particular form of hell. And Claude provided just a huge amount of emotional support for me and my husband as we were going through this process with my son, but gave us a lot of practical, actionable tips and even helped provide ideas for games and toys and things to do to keep my son occupied when we were potty training him. That is a really good tip. Thank you. Sure. So if Claude could help with one part of your day, what would it help with? Well, other than potty training, I really think that Claude can be an incredible
[0:02:54] They're an incredible tool and partner for productivity. And so much of what I use Claude for in my day to day is really helping me as a leader become better at understanding how to manage well in an organization that's so rapidly scaling. Anthropic was about 300 people a year ago. It's about 1,000 people now. And I've really helped to sort of page a lot of context around the company, my reports, things that I'm working on as a leader. And Claude has been a really great coach and tool for me to partner with.
[0:03:23] Very cool. I completely agree. So in five years' time, what's one thing we'll be doing with AI that feels completely unimaginable today? This is obviously a little bit more aspirational, but we really believe at Anthropic that many of the AI-powered tools that are being developed will be able to help with everything from
[0:03:44] Biomedical science and research, so helping scientists and pharmaceutical companies to develop better, more effective, safe vaccines, helping to cure many diseases that we have been struggling with over the course of human history. So we have really big aspirations at Anthropic, and we believe everything from really, really fundamental things about how to help make humans live and flourish better, all the way down to really helping to automate many of the tasks that we all have to do day to day.
[0:04:12] Very exciting. Yeah, they deserve a clap for sure. It has been such a joy collaborating with you as we have brought Canva code to life. The care you bring to partnering and iterating on your products has been really impressive and has really resulted in such an extraordinary product. So I'd love to know, what does the process look like for you all gathering the feedback and then deciding what you build next?
[0:04:36] Anthropic is a little bit unusual in the AI field in that we primarily build for businesses. And so something that's very important to us is really understanding very deeply what are the needs of our customers and our users. And I think we really have this in common with Canva, where we spend a lot of time with our customers of all different stages, really understanding what are the features that will make AI more powerful and transformative within your business. And also, what are the core qualities
[0:05:05] So there are a lot of really great similarities between Canva and Anthropic and our values. And some of the values that I love from Anthropic are
[0:05:33] To act for the global good and to put the mission first. What do these values mean and how do you bring those values to life and lead Anthropic with them? I think Anthropic was really founded on this belief that you can develop transformative AI technologies in a way that is
[0:05:50] Good and safe and beneficial for people everywhere. And we made a lot of decisions early on in the business. We incorporated as a public benefit corporation because we had this very strong foundational belief that AI was something that would transform how people work and operate together. And we wanted to make sure that we were doing it right.
[0:06:10] And this has led to a lot of different things that we do in the day-to-day of a company. So we have an effort called Beneficial Deployments where we really think about how do we want to help give back to so many different parts of the world, whether that's in education, developing world, health care, and life sciences. We also have something called a Responsible Scaling Policy that is a set of commitments that explains transparently what are the ways that we are working to ensure that the technology that we're developing is safe.
[0:06:39] However we push a new model to market. And I think there's a lot of other examples, but I think the sort of core foundation for me is that we work very hard every day when we think about making trade-offs within the business. How do we make sure that we live true to these values? And how do we make sure that we live up to the public benefit mission under which we're incorporated? So incredibly important at this stage in human history to have those front and center.
[0:07:05] Could you share an example of how these values have really helped to guide you through a challenging decision personally?
[0:07:11] Yeah, so I think the responsible scaling policy, I think, has been a very interesting experience for Anthropic. And really the way that we sort of designed it is we want to make it practical and usable for ourselves, but also something that is exportable to other businesses and to other companies. We really view ourselves as trying to create a foundation of ethical guidelines around how AI can and should be used. And the process of developing that is very complicated.
[0:07:40] Because it's not just sort of current focused, it's also forward looking, right? So we say, if models would be capable of doing the following things, how would we want people to use them? And how would we want to make sure that as a model developer, we're ensuring that the systems we're giving to other people are safe? And we've had sort of a lot of debate internally about what are the right ways to do that.
[0:08:00] And so much of what I think we've learned through that process is we have to be iterative as we go because the technology itself is developing and changing so rapidly that we want to be both committed but also flexible so that when we get new information, we can update accordingly. I love that approach of thinking about if AI could do anything in the future, how would we want it to be acting? It's such an important way to be thinking about it. So thank you very much for doing the work that you do.
[0:08:28] We talked today a lot about bringing productivity and creativity together. And is there an AI-powered habit or tool that you found has made you more productive personally?
[0:08:37] Yeah, so I think I really resonated with so much of the work that we've done together on Canva code. And this is because I myself am not a developer. And I think, you know, I heard you and your team stand up here earlier and talk about just the powerful unlock that comes from being able to develop something yourself if you're not somebody who knows how to code. And, you know...
[0:09:02] I talked about this and we built this into our product. I knew all of that. But the first time that I actually sat down with Claude and was like, help me build a website, it was incredibly empowering to be able to access this way of creating and developing that I had felt was just inaccessible to me. I always thought, wow, there's these really brilliant people in the world and they're Python developers or they know HTML. It's just not a skill I have. And I think the ability for Claude to really help up level my own skills and abilities
[0:09:31] By really unlocking this core fundamental skill has applied across a wide range of things for me, right? Everything from building interactive games and tools with my son to helping to develop a website. And I think there's an incredible amount of potential in that for so many people around the world regardless of their age, education, status, abilities, right?
[0:09:52] Couldn't agree more. It's completely changed the way we do product development at Canva as well. It's kind of feel like going, have you all done snorkeling before? And you go...
[0:10:02] You go scuba diving and all of a sudden the world feels like it's in 3D. It feels like the same for the product development process. It used to always be in mock-ups. And then all of a sudden you're in this 3D interactive world and you can take an idea and turn it into a design that's interactive in just a few seconds. I have terrible vision and the first time I put glasses on, I felt like I had the same experience. I was like, wow, people can see birds in the sky. This is amazing. Exactly that vibe, yes.
[0:10:28] So you've also spoken a lot about AI as a tool for collective benefit, not just individual efficiency. What are some of the meaningful or mission-driven uses of AI that you hope become more mainstream? So I talked a little bit before about this idea around biological sciences and I think the incredible potential that AI could unlock there.
[0:10:50] I also think there is a huge amount of opportunity in the space of education. And I think in particular, something that is very exciting to me is a tool like Claude's ability to really meet different people where they are. Something that I think is very, very powerful about AI as a tutor is the patience, the empathy, and the flexibility of an AI-powered tool is very different, and I think scales sort of a little bit differently than a lot of previous versions.
[0:11:19] And something that I think
[0:11:21] We've really seen is that Claude is very good at adapting to the individual person that it is talking to or working with. And I think whether you are an educator or an administrator or a student, there's an incredible amount of opportunity ahead of us for Claude to really, or tools like Claude, to really enable and empower individuals to learn about a wide range of things, right? Maybe somebody like me actually wants to be a developer and says, I just have never felt that someone would sit down with me and sort of tailor.
[0:11:50] And I think the amount of unlock that could be provided in the education space for people, again, of any age, of sort of any background, with any different learning styles, to really sit down and say, I want to become good at something. How can I create a plan to do that? Like, what should I read? And also, how can you help walk me through places where I'm getting challenged or stuck or demoralized?
[0:12:16] It is such an extraordinary opportunity we have, I guess, as humanity right now to ensure that every student everywhere in the world has access to that high quality education. I'm very excited to be partnering with you on that journey ahead as well. So William Gibson famously said, the future is already here. It's just not evenly distributed. And AI has the potential to change that. How do we ensure it benefits everyone and not just a select few?
[0:12:45] Something we think about a lot at Anthropic is really how do we ensure that the kind of distribution of benefits from AI technology are felt more broadly around the world.
[0:12:57] And I have a background before I worked in AI actually in global development. And so this is a topic that feels very near and dear to my heart personally. And something we're exploring this year is how do we really help ensure that AI technology is accessible really around the world, right? And Anthropic thinks of itself as a global company. And while, of course, we're starting to build in places like the US, we're thinking very critically about how do we make sure that access to this technology is something that
[0:13:27] All of that is really...
[0:13:29] Available to people regardless of where they are in the world. And I don't necessarily think there's easy answers, but something we feel very strongly about is that without active effort, we worry that AI will be concentrated just in the particular places that all new technological breakthroughs have been. And I think there's a real opportunity, not just for individual companies, but for civil society and governments more broadly to really think about how do we make sure that this is technology that helps people flourish regardless of where they live.
[0:13:57] I could not agree more. We talked a lot in the years gone by about the digital divide, but the AI divide, I think, is going to be extraordinarily important for us to be actively developing to prevent. So creativity and technology have always shaped each other from the printing press to digital design. AI is certainly that next wave, but where will it take us? Looking ahead, what excites you most about how it will impact the future of AI in creative industries?
[0:14:24] So we recently ran this very interesting sort of study and report, and we published it. It's something called the Economic Index. And what we found is that, at least today, and this is a little bit unexpected for us,
[0:14:39] The majority of AI use cases are what is called augmentative. So rather than saying an AI is replacing a customer service agent or it's replacing a developer, really what we're seeing is that Claude and tools like it are helping humans to augment things that they are already doing, creative work in particular that they want to do. And I think there's an incredible future ahead of us, if we can kind of get this right, where AI is this incredible force multiplier.
[0:15:09] That helps you to figure out what are the things that I really want to develop or build? What's important to me? What are my values?
[0:15:16] How do I use this tool to kind of help accentuate the things that I'm really good at and fill in the skills in places where I need help? And I imagine this world where AI is essentially helping to empower people to live the lives that they want to live, to pursue things that they might be interested in, and to really just help accelerate for their own development and learning. How do I want to get better at maybe a skill that I haven't had before? And I think there's a lot of opportunity, both in the creative space as well as in many, many other parts of the economy.
[0:15:44] I could not agree more. Do you see any huge breakthroughs on the horizon that we all should be aware of today? Something that I think has been really interesting for us is...
[0:15:57] We don't necessarily always know in advance what the kind of next model breakthrough is going to be. And today, Claude is particularly known as being this very powerful developer tool. And that Claude is so great at writing code. But if you had asked us maybe a year ago, do you think that that's going to be the kind of breakout use case for Claude? I'm not necessarily sure that we 100% predicted it. We were thinking, hey, it's a possibility that Claude could be a great partner for developers.
[0:16:27] In some interesting ways, the models themselves actually sometimes can kind of guide the conversation. They sort of show this emergent behavior as being very, very good at something. And so we have some inklings of what might be coming next, but it's a little bit more caveated than normal. I definitely think there's going to be continued advancements in coding and related disciplines, things like data analysis, financial analysis. But I also imagine that we're going to see really an uptake in the use of Claude for things like copywriting, support with marketing.
[0:16:56] Our internal sales team uses Claude a lot. And I really think there's this kind of interesting interplay between what the models themselves can do and then what are the products and features that a company chooses to build to really support the businesses that are developing alongside this technology. Very cool.
[0:17:18] As discussed, I think one of the most exciting frontiers of AI is education. And as we develop Canva code, we work with teachers to explore how it could be used in classrooms and the possibilities for personalized learning seemed completely endless. What are some of the most exciting ways that you see it shaping education? Yeah, well, some of what we talked about before, I think is absolutely top of mind for me. I think really this concept of having a personalized tutor,
[0:17:48] Context for you over a long period of time I think could be really valuable. And I think that that could be incorporated into a lot of different products. If you imagine a user who has a long history with Canva, the sort of
[0:18:02] The corpus and body of knowledge that you might be able to have for an individual, I think is like quite staggering, right? These models have this ability to ingest a huge amount of context and information. And so it's almost like you could have this personalized tutor kind of for life, right? If you think about in the audience, maybe a very, very influential teacher or professor or mentor that you had, maybe you had a relationship with them for a year, or if you're really lucky, maybe five years, maybe they've sort of stayed in your life for 20.
[0:18:31] I think that could be a very powerful personalized tool to help with
[0:18:50] You know, pursuits that range from traditional schooling to creative endeavors, helping you to become a better pianist or a better writer or a better teacher. And I think, again, obviously very far off, but I think there is an incredible power in the sort of personalization that can come from an AI getting to sort of know you over a period of time. That's a really beautiful way of putting it. Yeah, really stretching out that relationship is extraordinarily powerful.
[0:19:18] So we are almost out of time, but for anyone still figuring out how to make AI part of their own work, where is a good place to start? I think what I would say here is...
[0:19:30] Don't feel afraid of or intimidated by the technology. And this is something that I think is very common, right? This sort of idea of, you know, particularly if you don't have a technical background, which again, I don't, it can feel almost a little bit overwhelming to sort of say like, where do I get started? And I think my number one piece of advice is find a thing that you are really curious about or that you want to learn or that you're struggling with a challenge in your daily life. And just try one of the AI tools out there and say, how can this
[0:19:59] Help me. How can I learn from this? And also, what does it feel like to use the technology? It's evolving and developing so rapidly that I really think even a sort of passing knowledge today will be incredibly valuable in years to come as these AI tools just become more powerful.
[0:20:15] Could not agree more. Great advice. So thank you so much for making time for us all to join us here. And thank you so much for being an incredible partner. We love working with you and your team and can't wait to see what you're going to be doing over the months and years to come. Please join me in thanking Daniela. Thank you.