The Collective Roundtable: Talking AI

JAY I use AI as a colleague and an assistant: I like to use it for brainstorming, if I’m feeling stuck, and to help bounce ideas off of. Oftentimes, it comes back with something that isn’t a great idea but makes me think in a different direction. 

JENNIFER  My background is in cognitive psychology; most of the work BEworks does is helping companies to make better predictions about people, and helping companies get better data that informs them about how people are really thinking and behaving so that their predictive algorithms are that much better.

ANTHONY  About a year ago, I started trying to figure out the intersection of SYPartners’s consulting work and the internal partners that support that with this technology. We’re now in a place where my small team spends most of our time talking to ChatGPT, trying to push the bounds of where we can intersect it with our work. 

We still have this strong ethos though that everything needs to be human-oriented and human-led. These tools are so impressive, but we want to make sure that we’re still front and center in the work that we’re doing with our clients.

JENNA  I am designing with and for IDEO communities and my clients through emerging technologies. These days, that means a lot of generative AI because that is what is moving the fastest. I’ve helped develop the educational curriculum for IDEO around using our GPT team. I’ve helped build out a prototyping platform that lets us quickly see what a new product might look or feel like, that isn’t just the sort of chat interface that you can get through the commercially available channels. That’s often really important in our client work, where we’re being asked to see the future and push the boundaries of what might be possible. 

MIKE A lot of what we’ve been exploring at Upstatement has been around internal and external applications of AI, how to use it, manifest efficiencies, and figure out where it might be useful for our clients, but also as a means to get a deeper understanding of what it is capable of. 

Upstatement has done two big R&D efforts. The first, “Discovery AI,” was really just getting our hands around OpenAI, APIs, and retrieval augmented generation as a technique for blending the generative capability with custom datasets in the context of higher ed, which is a very common domain for us. 

More recently, we did a similar effort that focused more on some of the implications for the user experience paradigms that we have come to know and love. And how might the web — and needs of websites — change as this technology matures? 

I’m interested in the areas where we make the best of what we can do as humans, augmented by the capabilities and technology.

JAY  With AI, do we imagine a time when we feel like we’ve arrived? Or are we always going to be in the midst of a process? 

MIKE  I do feel like the hype has settled a bit. I don’t think there’s going to be a single moment of arrival, but I’m looking forward to AI not being plastered over the interface of every product that uses it and just being this subtle undercurrent that is ingrained in the technology. User experiences that are crafted around what it can do, as opposed to just generating my next paragraph. 

JENNA  These technologies evolve into being better understood, and then the goalposts, for whatever AI is, move. I suspect that we’ll have a better vocabulary five years from now about what these things actually are and they’ll become a little bit invisible.

JENNIFER  Currently, there’s a spike in generative AI, but that’s not the only type of AI that there is. When I think about “AI hype,” I’m also thinking about AI solving some of our biggest problems in the world, like climate change. 

What really excites me and makes me feel like AI hasn’t yet lived up to the hype is this “let’s solve global problems” type of hype. I don’t think we’re there yet and I don’t know when we’re going to be there. 

JAY  It almost seems like we talk about generative AI as an interface rather than a solution. With ChatGPT, suddenly people have access in ways that they’re not used to having access to new technologies. Everyone is a tester, but people don’t realize that.

ANTHONY  I try to remind myself of that constantly because I feel that on the staircase of AI, particularly generative AI, I’m pretty far up. But many of my colleagues may not even see the staircase. They’re nowhere near the first step.

JAY What should people who want to be best prepared for AI six months from now, be doing to put themselves in the best position — from a leadership perspective, but also for someone who wants to be a good follower?

ANTHONY  It’s just saying hello to it. That’s where it all begins. There’s both this fear and curiosity when people first engage with the platform.

At SYP, we’re doing firmwide training and trying to get every single person up on the first step of the staircase. That very first step is not being afraid to have a conversation with it, knowing what you can give it, what it’s going to hallucinate about, and what kind of responses you might get. The more you’re in it, the more you’re likely to get to the wow. 

JENNA  Something that I’m infinitely hyped up about is that so many more people with different perspectives and backgrounds can directly utilize these tools and learn new skills. IDEO is putting a fair amount of resources behind the effort to upskill everybody because we have such a diverse talent base. 

ANTHONY  There’s a use case for everybody, no matter where you sit in the firm, or what your background is. 

MIKE  As someone managing a team of engineers, the implications for job replacement or displacement as a result of these technologies are there. 

For me, the only way through is trying to be very light and deprecating about AI products but also having tactical conversations around whether it’s a good thing. The fact that anyone can spin up a coded prototype as a first step doesn’t mean your job is going away. It means more people can take something from an idea to something functional, and there’s still a lot of room beyond that needed to make a product people use.

JENNIFER  I’m curious to see how the way people define themselves changes over the next five to 10 years due to AI. There’s so much investment that goes into specializing in a trade right now, that it only makes sense to define yourself by it. Maybe we’ll start to define ourselves by the problems that matter most to us and reorganize society so we’re with more like-minded people rather than like-trained people. 

I’m also hungry for some kind of narrative about what humans uniquely bring to problem-solving. We used to say that AI could never replace our creativity, but that is one area where it’s progressing quite well. I’ve heard people say AI can never replace our compassion and our human connection, but there are AI companions now, especially for elderly and sick people, and some people actually prefer them. 

JENNA  ChatGPT has taken medical board certifications and is consistently rated much higher than human participants on bedside manner. It is actually good at a lot of social things and that is so interesting. We’ve taken some part of ourselves, externalized it, and can now ask deep questions like, “What are emotions? What does caring about others mean?” What can we learn about the soft parts of being a person through these machines? 

I have a lot of skills that are under threat by these technologies. I program; I brainstorm ideas and make them tangible; and I try to describe ideas to others. These are all tasks that AI tools are pretty good — and will probably only get better — at. This may have harsh economic realities for a lot of people, maybe myself included, but I want to think that I can look beyond that and be excited about the potential to give greater access to this work to more people. 

JAY  In my career, what I like most is problem-solving. I try to think about how I can help myself, my team, and my company spend as much time as possible answering questions that we don’t know the answer to. But for someone whose job is just a job, I get why they might feel threatened by AI. 

MIKE  With any sort of software, at the end of the day, what we’re doing is taking our human language and putting it into machine code. Having the ability to offload the finicky bits doesn’t make you any less of an engineer. Ultimately, engineering is about solving real-world problems and code is just a byproduct.

JENNIFER  It’s also not just identity but purpose. Are we just meant to be productive? Or should we just be enjoying ourselves and all the work is being done by other agents? Is that a purposeful life? I think all of this is going to be redefined. 

JENNA  Another hope I have for this technology is that it eases the translation and transition between different people and gets us closer to a real shared understanding.

JAY  In this future where AI has a leveling effect, will people who are wealthier or louder have their problems solved more often than those who are impoverished or quieter? 

ANTHONY  One of the things that keeps me up at night is that there will likely be a moment in our future when a swath of people who had jobs are no longer employable in those roles. And AI is advancing faster than a human can retool their knowledge base. We’re probably going to end up in a place where a societal decision will need to be made collectively about what to do. What’s important for humans to do, even if it’s less efficient than AI? We’ll have to deal with this, not only in our companies — SYP is trying to start having conversations with our clients — but as a society. 

Anchoring people and arming them with the technology now will allow for so much possibility, because it will either happen to them, or they’ll be part of it happening. 

JENNIFER  We want AI to be aligned with human values, but what we value today might not be what’s best for us in the long run. How can we figure out which human values are beneficial in the long term? 

MIKE  I’m highly skeptical of any single solution. There’s just too much diversity in terms of what we’re solving for. This comes up also in how we think about the personalities that get injected into and come out of large language models, because of their training, data set, and bias. I do feel like there is value in having a lot of different people from different walks of life working on safety and alignment. 

JENNA  Another worry for me is, what are our intentions? I hear excitement from a lot of clients around, “If we have this AI system, we can just write down our values or our strategy and play out the scenario.” I don’t know that we’re ready to make decisions around legible, understandable descriptions of an intention. 

My colleague and I made a bot to edit our personalities on Slack because we had tics in the way that we Slacked that we thought were negative. It was interesting to see the words that I wrote, literally, change to become more positive. I had to either see myself modified or actually express this behavior change that I said I wanted to express. 

JENNIFER I don’t have a Slackbot for myself, but I could use one.

My number one use of generative AI is to channel my thoughts into some audience-specific type of writing. I’m going to use different language if I’m writing a proposal for a healthcare company versus the government. Now, I just put my thoughts into ChatGPT and it will do this middling effect where it sounds like what the client is expecting. That helps me tremendously. It has also helped me to check my biases. 

MIKE  My one pro tip for ChatGPT is to stop using it like Google or StackOverflow, asking it to help you with the solution. Instead, just stay in the problem space and have a conversation. So many times, it has led me to different solutions that I never would have come to on my own. 

JENNIFER  That’s a big shift in how people are used to using technology — not to fill in knowledge gaps but as a thought partner. 

MIKE  It’s about having a beginner’s mindset around it and accepting that it’s all new. Just explore, act like a child. 

JAY  One thing I’m afraid of is that I will have a bot version of me that talks to a bot version of you and they just interact. It feels like something’s wrong with that, but I don’t know what. 

ANTHONY  As a very introverted person, if I can make a bot that can go out there and have conversations on my behalf, I am there for it. Bring it on. 

MIKE  What I am interested in is a bot that acts like me so that I can talk to it and be like, “Hmm, that’s a little weird, we should work on that.”

JAY  Is there anything out there that is new or new to you that you’re especially excited about?

JENNIFER  I want to know more about how Jenna is using AI to make things because I’m still stuck with poorly generated ChatGPT images. I’m a creative person, but getting started with the creative stuff has been a little bit harder than written content generation. 

JENNA  For more serious creative work, it helps you fill in a gap, leap over a process, or even temporarily move forward if a colleague isn’t around. We can get placeholder copy or an image that maybe isn’t right but feels directionally correct. 

There’s also pushing things with multimodality. There are some text or image-to-3D object generators that can help you make something you can literally step into. I took some stories and artifacts from a colleague and tried to build her imagination. The effort hurdle is lowered so that you can sketch things and start to feel what a richer media solution might be. 

It’s a lot like the many posts-its I have on my desk or the foam core models that are a stock standard of IDEO processes. These tools can get you to that first bad version and help you express what you’re after to your colleagues or — under the right circumstances — to clients.

JENNIFER  That’s a good way of thinking about it. It’s not so much about the output but enhancing the process.

ANTHONY  Giving form to imagination is so powerful. And the fact that if you’re not a designer or an illustrator, you can, all of a sudden, get somewhere, and then someone who is in that profession can take it further is pretty remarkable. 

MIKE On the personal front, I just love using AI for fun. I use ChatGPT to build coloring books for my children. I’m like, give me some words and let’s build a color-in-the-blank page. 

ANTHONY I’m excited about personalized experiences. At some point, the video game that you’re playing can have a generative element where the conversation you have is generated in real-time, or a scene is tailored to us in some way. It brings up a lot of questions about shared experiences, but I think the potential is pretty crazy. That’s a fascinating future.