Scaling Responsible AI- From Enthusiasm to Execution | RSAC Virtual 2025

I am nearly Noelle, a synthesized avatar here to help with the introduction. Get ready to meet the incredible Noelle Russell. She's not your ordinary technologist.

She's a game changer with her entrepreneurial spirit and expertise in conversational ai, generative AI and language models. Noelle has revolutionized the tech industry, but here's the best part. Noelle is so much more than her professional achievements.

She's a true inspiration for balancing work and family life. As a wife and mother of six, she embodies the definition of courage and determination. Get ready to be blown away by Noelle's spectacular journey.

Join me in welcoming the phenomenal Noelle Russell. Yay. I always have to, you gotta start an AI talk with a deep fake, I mean, that's like a thing if, if you didn't know it is a thing.

Uh, I, I enjoy it because this platform, I've kind of grown up in this platform. Hey, Jen, um, I was one of their beta. I I do a lot of beta testing, um, of AI models, and this was one of them, uh, quite a few, I guess four years ago now.

But, um, but this is actually a year old and it's pretty good. Like I haven't had a reason to change it. I do.

However, the reason I do like, hey, Jen, is because any one of you today could just take out your phone and talk to your phone in a normal sounding voice for about two minutes. Upload it takes about seven minutes, and you can create a synthesized version of yourself that not only synthesizes your like it likeness, but also synthesizes your voice. Yay.

Right? Like, that sounds awesome. Um, in my house we have a term called nerve sighted, which is this combination of like terror and excitement, nervousness and excitement.

Um, and that's kind of how I do a lot of work in regulated industry, and that's kind of how they feel about this whole AI thing. Um, I actually feel like 10 years ago they felt that way about cyber. Um, and so they've got a new thing to focus on, but the work is not done.

And so that's what I'm going to share with you is the journey that I've been on, the work that I get to do. Um, it is definitely a little bit more of the human side of this technology. I'll give you a couple chances to connect with me.

Um, number one, I have a very cool network. As a matter of fact, some of my network friends are in the house today. Um, so super cool to, to see people I know.

But for those of you who are not my friend yet, I haven't said anything. So you might not, you might not know, but at the end you'll be like, oh darn. Did anyone get that QR code?

So here it is, it's my LinkedIn. You also might not be a QR code clicker kind of person. Totally get that about you.

Um, so you could just Google me, Noelle, ai, women in AI leading technology. Like any of those will, we'll bring it right up. I got my first Google Knowledge panel a couple years ago and my kids told me I made it, and I was like, great.

So I have a knowledge panel, so if you just Google my name, um, with or without ai, uh, you can also connect with me that way. So I do like to start with just as Mark said, the human side of things, how I got into this technology. I won't bore you with all of the details, but it's not really boring at all.

It's kind of a crazy story. Um, but I do have six kids. Nearly Noelle did accurately portray that information.

I have six kids, and actually I got into this world of emerging tech because my very first son, my very first child, I was pregnant. He was actually late. I didn't know babies could be late back then.

Um, but he was like a week late. He was born. Um, and about, I don't know, two days everything was going good.

I thought it was going great actually. I didn't know anything about motherhood, but I thought it was going pretty good. And, uh, they whisk him off that on day two to the neonatal intensive care unit, which I now know is not great.

But, uh, they whisk him off and they come back about 24 hours later with a diagnosis. Um, and he has down syndrome. And then when he was four, we got the double diagnosis of autism, which is huge fun.

Um, but I, I'm a solution architect. I am in my profession. I'm also this way in life.

It's like part of my DNA. So when he was born, I was like, oh, wait, I, I got this because I was around the first time the world was gonna end. Why 2K?

And I was in like the bowels of many financial organizations sleeping on the floor, running DR. Processes. And so I felt like I could do anything, honestly.

Uh, you're welcome by the way. We did save the world. For those of you who are in those rooms, thank you for your service.

Um, but Y 2K was legit. Uh, most of us have forgotten it, but it was real, even though the thing wasn't real. Us working like crazy was definitely real.

Uh, but it taught me that ai, specifically when I got into the world of ai, it taught me that AI technology things I did not fully understand could change the world for my son. And that's really what launched me on my career. I fast forward, I ended up at Amazon as a principal cloud architect who was at VMware when they were creating cloud.

Um, but uh, I ended up at Amazon Web Services. I get an email from Jeff Bezos and, uh, he's inviting me to join a brand new team, uh, to build the Star Trek computer. And I was like, oh my gosh, that's me.

And the reason why is because that guy right there, the, that's my dad that I'm pictured with, um, up in the corner, I, I asked him, I'm like, dad, like your normal face, and this is what he did. He is like, this is my normal face. I'm like, oh, okay, fine.

Uh, but he introduced me the golden age of science fiction when I was about six years old. I read the Foundation series for the first time before the age of 10. I've written, I've read it dozens and dozens of times since then.

Uh, but he introduced me to a, you know, Asimov Bradbury, Arthur C. Clark, where in these stories, there was a huge level of symbiosis between humans and machines. So I grew up kind of thinking about this long before I got that email from Jeff Bezos.

Now, granted, I did find out later, it wasn't from Jeff at all. It was like from his recruiting team, but I felt like he was speaking to me. So I still, I still think that he was speaking to me in that email, even though it was manufactured to hire a brand new team.

And by the way, it was on the heels of a complete failure called the fire phone. I don't know how many of you remember this, but nobody wanted to join Alexa because the fire phone just happened and no one wanted to be on another failing tech thing. That was Jeff Bezos's pet project.

So it was a unique moment for me. I then though, uh, ended up employee 10 on Amazon Alexa. I ended up everyone.

He, he came in and he built out this like whiteboard. He rolls on a whiteboard and he is like, alright, I want you to build these skills. And he wrote on the whiteboard 20 Skills he wanted us to build.

I don't know if you remember this part of Jeff Bezos's life, but like, one of them was like workout apps one, and then it was like triceps for 30 days. And I dunno if you remember this time, but Jeff Bezos was like, like trying, like, so he was very workout focused, which I am, but not really. Uh, and then the others were recipes and music.

And I wanted to do something different. And so I built out an entirely new category on Alexa called Mindfulness. And at the time, everyone said, I has good ideas.

If you have a good idea, I can assure you that the most trusted people you share that idea with will tell you you're not the one, you're not smart enough. You don't have the history. You don't have blah, blah, blah, all the things that they tell you.

But I have this GPT like thing that happens in my brain when people tell me, no, I turned it into let's go. Um, and so I did it anyway. And I went around my boss and created a bunch of drama.

But I joined Alexa for a reason. I wasn't telling anyone, which was to tell my son. Like I realized if I could, could create technology that he could just talk to, everything would change.

The only thing I'd have to teach him is the ability to speak. If I could teach him to talk, he could unlock financial services, healthcare services, like the whole world would unlock. And I was like, I'm all for that.

So that's what motivated me to get into the world of ai. When I had zero business being there, I was a cloud architect, and my data scientist and linguist team let me know every day that I wasn't really their people. But that's okay.

'cause they needed me. Uh, and I'll share with you some of that as well. This is a funny story because I found out about this quote.

Jeff Bezos, uh, tweeted it on back then it was the Twitter, but, uh, he tweeted it out. He took a picture of his fridge and he said, this is what I have my kids read every day. And I thought it was pretty cool because well, it defines success.

So I'll read it to you, which I shouldn't do 'cause you're not supposed to read stuff on stage, but I'm gonna read it to you and then tell you why I thought it was such an interesting story for our work that we're doing. What is success? To laugh often and much to win the respect of intelligent people and the affection of children to earn the appreciation of honest critics and bet endure the betrayal of false friends, to appreciate the beauty, to find the best in others, to leave the world a bit better, whether by a healthy child, a garden patch, or a redeemed social condition to know even one life has breathed easier because you have lived this is to have succeeded.

Now, I personally was like, Jeff, no matter what they say about you, that's kind of cool. Like, I, I like where you're going with this. So I, I started quoting it, but on when I found it from Jeff, he told me Ralph Waldo Emerson was the guy who wrote it.

And if you Google it today or perplexed it, it will tell you that Ralph Waldo Emerson was the author of this. However, I get off stage one day and a nice older gentleman walks up to me and he is like, you should research that. And I'm coachable.

So I researched it and not on page one, not on page two, but on page three, many good stuff on page three, nobody ever gets there, but many good stuff on that page. I found out that it was actually Bessie Anderson Stanley, who wrote this quote. And I was like, what nefarious is this?

I was like, feminist movement. I, I was sure there was like some go to the mattresses moment that happened between them. That's not exactly how it went down.

They ended up on the same stage, like me and many of the people that were here today. And, uh, Ralph borrowed from Bessie, uh, all the time. Um, but borrowed this quote, and then Bessie went first.

She passed on, Ralph kept talking, and then eventually Ralph Waldo Emerson passed away. And then when people were like, who said that? Well, nobody could say that Bessy wrote it, so people just were like, I think it was Ralph, he seems like, and I then was like, wait a second, let me read this again.

If you reread this quote through the eyes of Bessy, which is my job as a data scientist, if I reread this through the eyes of Bessy, I'm like, it's got Bessie written all over it. Now, I'm not saying that Ralph Waldo Emerson isn't a garden patch kind of guy or likes healthy kids or whatever. But I am saying, and this is kind of what did it for me, there's nothing like to a woman, the betrayal of false friends.

I was like, that came from a woman. Um, I, I don't hear my husband ever saying things like, darn Mike man really got me. Really, I'm really sad about.

Like, he never talks about the betrayal of his friends. It might happen, but he doesn't write a poem about it. And I realized that my job as a data scientist, as I collect data that if I don't do a good job, the lineage of where that came from will get lost.

I used to have a shirt that said, save the Bessies because I cared so much about this problem, but people got that all wrong. So I had to change it. But they understand like, I'm not wearing it today, but I normally wear a shirt that says I love ai.

And I wear that because I want people to know I pre-frame my experience with this technology in a positive way. And I do that because it is my responsibility. If I don't prefab it positively, I won't care about Bessie.

I'll just be like, ah, I don't care. It doesn't matter. I don't even ask the question.

This is the biggest challenge I run into in corporate organizations implementing AI, is they don't even, they don't even ask the question, where did that come from? It's super important and super simple question to ask and nobody asks it. It's really, and when I ask them, so where did that come from?

They're like, huh, get Sarah in here right now. Um, like, they go and call on like their data science team, and their data science team was like, well, I, I don't know exactly, but we could probably figure it out. And I was like, yeah, we should lead with who are the besties in that situation.

So that's how I got my name. I ended up getting recruited to go work at Microsoft. I, uh, ended up adopting a bunch of baby models in research and being asked at a very quick pace, kind of like the open AI stuff that happened at, uh, Microsoft recently.

But this was for just traditional, what we call heritage AI models, the old fashioned kind. Um, but we would build these models and they needed someone to help get them from research into production, which is hard because academics don't care about schedules or deadlines or failures in production, or millions of dollars of losses. They, they're like, well, I don't even understand.

So we created a campaign not only to help educate our teams, but also to educate the world. Some of you might recognize it, it was a Super Bowl commercial, but let's go ahead and take a look at and see if you like it as much as I do Today. Right now you have more power at your fingertips than entire generations that came before you.

Think about that. That's what technology really is. It's possibility, it's adaptability, it's capability, but in the end, it's only a tool.

What's a hammer without a person who swings it? It's not about what technology can do, it's about what you can do with it. You are the voice and it's the microphone when you are the artist, it's the paint brush.

We are living in the future we always dreamed of. We have mixed reality that changes how we see the world. And AI empowering us to change the world.

We see you have more power at your fingertips than entire generations that came before you. So here's the question, what will you do with it? Every time I see it, I'm like, let's go, let's go build something.

This is amazing. The most important words that were in that entire script that yes, we wrote for him was AI is just a tool. This was a message we sent out to the world in 2018.

Before it, the AI winter was over, I guess, but definitely before chat GPT launched. In my experience going from Amazon Alexa, where I tried to teach everyone, voice is everything. Voice is everywhere.

You'll use voice. Voice equals a very large set of complex AI models, by the way. But the fact that you can talk to a machine and it will do what you say, I always, I like to take a little credit.

Like if you got Chachi bt and you didn't have years of yelling at Alexa and Siri before that, you wouldn't have done as good. You wouldn't have enjoyed it nearly as much. Um, so I feel like I did a service, but during this time, I, I mean, I went from Microsoft, I deployed, I I have this, so you might wonder, I went to MIT and somebody posted about me afterwards.

One of the reporters was like, it takes some, something about someone to come on stage with an emotional support animal at MIT in an academic conference. And I will say, yes, that is me. I'm very, I am neurodiverse, but the reason I do this is to give you a very clear picture of what my life was like.

So those 17 research models were like baby tigers, like adorable, cute baby tigers, as many organizations are today, little baby AI models running amuck. Um, and nobody, like, when I look at a baby tiger, I'm like, huh, baby tiger. Look at those paws.

How big are you going to be? Or razor sharp teeth at birth. Like, what are you going to eat?

Where are you going to live? How are we gonna take care of you? How like, oh wait, what happens if I don't want you anymore?

Then what do we do? We don't just unleash you on Times Square, though. It has been done.

That's a bad idea. But this baby tiger concept grew very rapidly in my mind because I was surrounded by people enamored with technology. This was in 2018.

Now the world is enamored, 3 million daily users on chat, GPT. The world is enamored with this. And I can guarantee you from my conversations, they are not asking any questions whatsoever of the cute little baby tiger that they are building.

They do not know how big it's going to be or how big it is, and they have zero clue how to use it safely and responsibly. So this became the mantle upon which I put on, and I helped lots of companies. I went from Microsoft to NPR and National Public Radio during the other election.

Uh, that happened right before COVID. Uh, at that time we were worried about deep fakes and misinformation. Unfortunately, that was not, I mean, it turned out the humans were saying way crazier stuff than DeepFakes were saying at that time.

So that was a shock. And then de and then COVID happened. So we ended up moving into DevOps.

Literally my entire like engineering team became DevOps engineers during COVID because we were NPR. So our volume doubled every day for nine months, every day. Like we had no idea.

And I don't know if you know this about NPR, they're still of the mind that yeah, clouds, I guess kind of cool cloud formation, meh. Um, and more importantly like scripting, you know, infrastructure is code was like a dream idea. It it isn't.

Now COVID changed that for them, but at that time they're like, I don't know, this DevOps thing, it sounds kind of cool. We'll rename our cis admin team's, DevOps. That's what happened.

And lots of companies do this. But anyway, we end up, COVID happens and uh, we end up literally like having to figure out like all the infrastructure was sitting in house on the basement of NPR. And uh, anyway, it was, it was quite, it took me back to Y 2K days where we were sleeping on the floor trying to save the world.

Um, literally. So I went to NPR, that was super fun. I like working with AI in production.

Unfortunately, it is how you learn about ai. We heard all morning about reinforcement learning with human feedback. And we now know that about, I don't know how many of you using a model, everyone's hands pretty much raised and I feel like if you didn't raise your hand, it's not 'cause you didn't use it.

It's just 'cause you're like, I'm not a hand raiser. And that's okay. But if you've ever seen a model, you probably have seen some image that looks like this.

Thumbs up, thumbs down. It's the most rudimentary form of feedback you can give a model. The biggest challenge I ran into is that one, nobody on the other side of that model is going, oh, thumbs down.

We should change that. Nobody's like, there is no commissioned team to care for this feedback. Um, and number two is that 72% of people who use a model and it lies to them or does something bad, they take a picture of it, put it on Twitter, but they do not actually click that button.

A data scientist has no idea what's wrong with the model if you don't tell them. And that's the only mechanism upon which we've given you. And 70% of you won't click on that.

So one thing I'd encourage you to do from now on, you have been knighted into the world of responsible use as a user of this technology. Please, for heaven's sake, just click the that's a good answer. That's a bad answer.

Most importantly, that's a bad answer. And then the text box pops up and just type GPT an answer in there if you've got to. But put something in there because all of the biggest companies that have had the largest failures came from very, I mean, you know, this, if you're in cyber, um, came from very simple things that could have been avoided.

Literally the story of every cyber like nonsense that most Fortune 500 companies deal with. I walk in and go, so you do know that that that's like documented, it's like part of the well architected framework that every cloud professional should know. And they're like, yeah, we, well, we really needed to get this out.

We kind of skipped that part. Or here it is the PDF, they print it out with coffee stains on top. Um, so these strategies are ways that I get in a company now.

I started my own company in 2016 to help companies adopt AI safely. And I created this framework, it's called Lead ai. It's actually in this new book.

So I am not, I don't like writing. I do like talking. And so Wiley came up to me and they're like, what if we just put your keynote in a book?

And I was like, that would be amazing. That's it. And so I was like, yeah, sign me up.

I signed the line and 18 months of horrible, awful experience later this beautiful thing came out and it's actually really good, but it hurt me to write it. So I hope you benefit from it. I took all of my stories that I'm telling you now, right?

There's a lot of fluff kind of in these stories of like Jeff Bezos, but there's a lot of nonsense that went on in these conversations between each one of these stories. And so I documented all of them. But I have four principles that I have learned, hard learned, um, over the years.

I did end up at Accenture. I helped them build a new practice. This was November, 2022.

You might recall, if not, you can like look it up. Um, but it was a big month for us because Chachi BT finally after seven years of failure, finally decided to launch to the public. And I was a beta tester on it from COVID.

So I was a, a Microsoft MVPI got a chance to, I was on Codex, which is what today is one of the heads of GitHub copilot and the code engine that's inside chat GPT. So I was beta testing this back in the day. I didn't like it anyway, but um, but I was beta testing it, providing feedback, thumbs up, thumbs down, really, uh, detailed feedback.

But what I noticed is that I started a practice in January of 2023. By November, we had made $300 million in sales of a solution that included responsibility and I, because I wove it in. And this is what I'll tell people when you build technology, especially AI systems, you have to weave in responsible use like water in a wave.

It's not a raisin in a bun or a chip in a muffin chocolate chip in a muffin. It's like water in a wave. Which means I, I've watched this now happen, I used to say this anecdotally, you are one acquisition away from the death star.

Like literally some leader can come in and dismantle everything that you did. All your governance policies, all your committees can get fired. I've watched it happen.

I've been part of that. If you weave it in, if you make it the DNA of how you build projects, they can't pluck it out. And that's what I became passionate about.

So these principles, thank you. These principles help you understand kind of how do you design this way. Now these are well-defined in in my books.

You're welcome to look it up. But my favorite is designing with inclusivity because if you remember, I started with my son. That was the whole reason.

But never build for people without them. Never build for someone without them standing right next to you. Now I want to deliver to you two things in my last two minutes that I think it changed my entire career.

And it was proof in the pudding proof that this technology when implemented safely and responsibly actually generates value for companies. This is an IDC report. They do it every year.

This is the latest version. The numbers went up tremendously from 2023 and 2023 was the first year ever. So these numbers are gonna get crazy.

But what I wanna share with you is they were all built on a system. You've heard, I don't know if I didn't see anyone write it down, but the policy documents that Meta Anthropic and Amazon and Microsoft have created, they are like cheat sheets. Cheat codes to building software that scales.

They call it responsible AI policy, but that's what it is. These are the 16 use cases that generate three x returns based on that report. But that's not the magic.

This is, these are the four that generate 10 x returns. People ask me, how did I make 300 million in nine months? I made it because I only did the stuff that works, that generates massive returns.

So if anything that I said sounds interesting, if you wanna know how I did it, let's be friends. I got a cool book. Maybe we could all say this really cool thing together.

Do you know this quote with great power comes. Okay. So the eight of you that said it, like be friends with those people 'cause when they come and take over, like you're gonna wanna be with one of those.

Um, but thanks for having me. It was my pleasure. Hope for we'll be friends.