“I think that the people that I’ve worked with in my labs have really been proponents of new technology. When you are working for the nation, you want to be on the forefront of technology. You need to know how to use that technology before other nations know how to use it. So if it can give you that competitive edge … In academia, there’s still this sort of hesitancy of need[ing] to do it yourself. You need to go on the blackboard and write it out, and write out every single step: ‘you need to use 100 sheets of paper, and you need to know exactly how to do it: you can’t use a calculator and you can’t use AI. I had to do it that way. So you also have to do it that way’ … In academia, I do think that there is a bit more of a hesitancy to embrace these new technologies. What I like seeing is, at least in Computer Science, I know that there’s some champions there that are very much [saying] ‘publish or perish.’ You can use these Gen AI tools to brainstorm and come up with ideas.”
Kristen Hallas
Emerging Research Scientist & First-Generation Scholar
Connector / Strategist / Investigator“We don’t pause enough to look at our stories that we’re telling ourselves, and if we look at the stories that we’re telling ourselves, we can decide if this story fits … or needs revision. If I look objectively, and I write out the story of where I started versus where I am now, my STEM journey, [I think] I’m actually doing pretty good. Sometimes we just kind of tell stories in our heads, like, ‘I’m not good enough’ … I think going to conferences has helped me tremendously. Just these spaces that exist that tell people, ‘go to grad school. Don’t just go for your masters. Go for a PhD. Yes, you can do research. It’s okay. You do belong.'”
Designing for Growth
Human-driven growth
“I’m a first [generation] scholar … I loved school. I loved computers. I dove right into computers. I liked writing, I liked math, I liked computer science. In high school, I took all those sorts of classes. It’s a time to set the stage … I don’t know that people were [energetic] about STEM, encouraging people to go into STEM, or talking about STEM careers. I did everything that I could do. I doubled up on my math courses. I took a Computer Science [course] – my sister really encouraged me to do that. When it came time to pass the class, I didn’t have a career pathway to really move towards … [After working in a few fields, I decided,] if I can wear 10 hats in a row, I could probably go back to college and finish my degree. I [was] a non-traditional college student, in the sense that I stepped out of college a couple of times. I [decided] I [was] going to do research. I went to different institutions, did research projects, started working with national labs [and] research projects. That’s kind of how I landed here.“
AI-driven growth
“I became a campus ambassador for Perplexity AI for my institution this past spring, because I felt really passionate about how the tool was helping me. It helped me prepare for my [doctoral] comprehensive exam. When you have to understand a problem from start to finish, you can write it and you can drill it so many times, but I found that I could just walk through the problem on Gen AI and then understand where the gap of misunderstanding [is] coming from. It is iterative. You can ask AI, like, how to solve [the problem] … It’s bridging the gap between where the misunderstanding is [and you], so I don’t need the Generative AI for that problem anymore; I learned it. It’s not that you know that a human couldn’t have taught me that, but when you get to a certain level of your education, you’re more autonomous. It’s more on your shoulders. You’re becoming an expert for a reason: you know.“
AI vs. Human-driven growth
“I started doing research with Livermore Lab, and I learned that they are developing tools for us to use internally. If you have proprietary information and you don’t want to share that out on a cloud server, you can ask it internally. You’re kind of Googling it, but you’re using Gen AI. I’ve really held on to that analogy, especially as I talk about the benefits of Perplexity. Consider it more like a search engine than anything else. You use it responsibly. It’s a way to enhance your work. You’re just more efficient. There’s so much expected of us. So if we can be more efficient, then that leaves more time for the things that matter to us, like family and being able [to] just take a breather – to look at the clouds.“
Designing for Agency
Human-driven agency
“If you’re pregnant and you’re a PhD student, and you’re having to juggle the program and everything … I [realized] I couldn’t just keep existing as if I wasn’t pregnant. I didn’t want to spill the beans, but I realized at that point that I kind of had to just say, ‘hey, this is happening. I still want to work.’ I’m blowing it up so much in my head, because by the time that I shared it with the staff scientist, my co-advisor on my dissertation, he was just so excited for me. He said, ‘congratulations, it’s fine. You’re gonna be fine, and we’re gonna keep working on this.’“
AI-driven agency
“I’m a first time mom. I wish that I had the time to read the 500 plus page baby book that somebody gave me for my baby shower. I loved it. I had it on my list to read but what’s in my hand all the time? It’s this phone. Especially when I’m stuck in one place, pumping … I literally have my hands so full. All I really have room for is the phone. I ended up asking the majority of all of my questions about pregnancy and now, postpartum and development. Just like a search engine [I ask], why is the model giving me this answer? It’ll have, like, a little number next to it [to check sources], and I just tap on the number, and then it’s like, I go, and I’m looking, I’m like, okay, it is helping me find information so that I can continue to do my research.“
AI vs. Human-driven agency
“I think everyone can be like a steward in the way these models are constantly learning, and they require feedback from us. Give that feedback if it’s a good answer. If it’s bad, give the feedback ‘bad.’ You don’t need to [tell] the AI ‘please and thank you’ all the time. That can actually take up more computing power, because the AI, when it receives the input of the ‘please and the thank you,’ it’s actually having to do more work and produce more words, which is more energy on the systems. I remember reading an article about that. I [thought] wow, okay, maybe it’s better [to give] the straightforward answer. But like, it doesn’t mean that you still can’t interact with these systems in other ways. [Be] responsible, [be] cognizant.“
Designing for Belonging
Human-driven belonging
“For my undergrad thesis, I did a lot of work on imposter syndrome [with] women in STEM. I did a series of interviews, and was talking to different women about their STEM experiences. I think that that just helped me realize that we’re all just like going through it. Talking about our STEM experiences really helps us to create our space. There’s a really poignant quote in the work that I was doing: ‘we don’t pause enough to look at our stories that we’re telling ourselves, and if we look at the stories that we’re telling ourselves, we can decide does this story fit, or does the story sort of need, like a revision.’ If I look objectively, and I write out the story of where I started versus where I am now, my STEM journey, [I’ll think] I’m actually doing pretty good. Sometimes we just kind of tell stories in our heads, like, ‘I’m not good enough’ … I think going to conferences has helped me tremendously. Just these spaces that exist that tell people, ‘go to grad school. Don’t just go for your masters. Go for a PhD. Yes, you can do research. It’s okay. You do belong.’“
AI-driven belonging
“Has AI changed the way that I view human connection? I think so… like I think that in both good and bad ways. It’s made like a therapeutic tool. Tools that you would learn about in therapy [are] more accessible to be used in action. There’s one that I use a lot called ‘check the facts.’ [I will say], ‘I am having these emotions right now based on this situation that I’m perceiving.” So, how can I check the facts and [say] this is the fact of the situation, and this is the feeling. Does the fact match the feeling?’ I think that AI has become a tool for me to say this is my stream of consciousness – what’s going on? [It] help[s] me look at [the situation] more objectively … I think AI has even started to talk in [an encouraging] way. It does help in [difficult] moments, because it’s something that maybe I wouldn’t want to burden somebody with – what I’m going through in that moment. I just want to be able to power through it myself. AI is becoming [a reassuring] tool [telling] you, ‘you are going to be okay.’
[However], I have friends that will stay up all night talking to the AI. They’ll ask about astrology. They’ll ask it for relationship advice. I think you can start to have these parasocial relationships with AI that could be really dangerous, and they’re not a replacement for a human connection at all … On social media, I’ll be scrolling sometimes, and [it will ask], ‘do you want to have a chatbot conversation?’ No, I don’t want to talk to [AI]. I want to talk to [a] human because these models will tell you what you want to hear. I think that’s how they can get so addictive to people. It gets used to the information that you want to exchange with it all the time. The AI knows exactly what [we] want to hear. That’s very dangerous to me.“
AI vs. Human-driven belonging
“[As human scientists], we can’t possibly go through all of [our lab] combinations ourselves. It just takes too much time. Time is ticking, especially when you consider different environmental crises. [We can’t] tell the person that’s sick that [we] need more time. [We] want a solution now. [We] have robotic agents that assist with the experimentation process, but not just because there’s autonomy [but also] automation. Without instruction, the model is just sitting there doing nothing, or it’s only going to do one thing, and it’s not going to be flexible or or adaptable for complex reasoning. That’s the motivation of what I’m working on right now, different pieces of a model that contributes to that reality.“
Want more?
Check out other portraits or try your hand using AI with our prompt library.
