Your AI is only as responsible as you are

**Ryan Donovan** (0:00)
Solving difficult problems is what drives you. Raytheon is hiring software, systems and hardware engineers to take on some of today's most complex technical challenges. Explore opportunities at careers.rtx.com/raytheon.
Welcome to the Stack Overflow Podcast, a place to talk all things software and technology. I'm your host, Ryan Donovan, and today we're talking at Microsoft Build about Responsible AI. And my guest for that is Sarah Bird, who is Head of Product for Responsible AI. So welcome to the show, Sarah.

**Sarah Bird** (0:39)
Yeah, thanks for having me.

**Ryan Donovan** (0:40)
So before we get into the Responsible AI conversation, let's get to know you a little bit. How did you get started in software and technology?

**Sarah Bird** (0:48)
It's actually a good story that I, you know, share with a kind of a good lesson learned, which is when I was in fourth grade, two women in engineering came to my school and said, engineering was a great career and there weren't enough women in it. And, you know, at that time, I think my teacher and my parents said, you're good at math and science, you should be an engineer. And, you know, they had given like a pamphlet about women in engineering. And I decided then that, okay, I'll be an engineer and I guess never really thought about it after. And I'm so glad that I went down that path, but it's amazing to think if those people hadn't come to my school, or if the people around me hadn't thought to suggest it. You know, I might be doing something totally different today. So, like, small moments can make a big difference in someone's life.

**Ryan Donovan** (1:35)
I think that's a good lead in to the AI conversation because there are a lot of small moments in AI that can change people's lives. And a lot of the conversation today is about responsible AI, right? Let's start on the other side of it. What is irresponsible AI?

**Sarah Bird** (1:51)
You know, I think there's, like, many different ways that can manifest, right? And so, in a lot of ways, it's kind of easier to talk about how to do it responsibly.
It's like, what's a secure system versus an insecure system? But I think the thing about where we see AI being developed or used irresponsibly, it's often because there's a lack of intention, right? It's a powerful technology, and when you're developing with it, you need to understand it and you need to be thoughtful about how you're putting it to work and what you're putting it to work for. And then the same thing as a user, you know, the same tool might be totally appropriate for one use case and not for another. And so, I think that often when people are just experimenting and really not thinking first, that's when you see these kind of responsible AI mistakes happen most.

**Ryan Donovan** (2:48)
I have friends who are vehemently anti-AI. There's, you know, risks of social effects, there's market effects, power and hardware, and then there's sort of labor effects.
How can we sort of mitigate those and also the sort of trust risks of hallucinations? How can we sort of be responsible with this whole large bucket of problems?

**Sarah Bird** (3:12)
Yeah, it's a real challenge today because I think as the technology itself has become, let's say, more visible to people, right? Because with generative AI, it speaks human language and so it feels much more personified. And so we're now connecting into these deeper fears and like the connection with science fiction, for example. So I completely understand the fears and we need to address this and make sure that people feel in control and they understand the technology.
But at the same time, it's sad because it's like AI is also not just one thing, right? It's a huge diversity of tools. And if we look in my work, for example, being able to use AI to find threats and attackers faster in security to secure our world's most important systems, like that's a huge breakthrough that we're having because of AI. And so, we don't want to look at the technology just monolithically and say I'm against AI or I'm for AI. It's about where you actually use the tool and how you use the tool. And so, the whole point of responsible AI is exactly to address these concerns and talk about what do we need to do to build the technology in such a way that people can trust it and that we can control it and govern it. And what do we need to do when we're deploying and using technology in a way so that it will do what society wants it to be doing for it, and not these other things, right? And so, it's a very broad field that is looking at exactly this point of how do we build and use the technology in a way that's in service of people and not technology for technology's sake.