I Participated in Anthropic’s “Your Thoughts on AI” Interview: How It Went

Yes! Let’s begin.

Claude helped me think through a communication plan to keep people in my organization informed in a way that matches our values, but also doesn’t add drag to people’s days. It was a helpful session where I collaborated with the tool.

I am new to vintage auto repair and I used a Claude Project with documents about a vintage car I recently purchased. During the repair process, I shared images of what I was seeing in the hood, to diagnose a problem, I was installing a starter and it was a repair I had never done before. Claude helped me check my work and identify parts of the car that the shop manual did not detail very well. The repair went well and it was helpful to have feedback while I was working on the project. It was an unnerving experience because I was afraid I would break the car, but Claude plus other resources I found on the web and the shop manual guided me through and it made me a better mechanic. Now what I work on the car, I am way more confident and I understand how it works without having to use the tools.

It underscored the importance of intent when using AI tools. If I am intent on using tools to improve my knowledge and skills, then AI tools actually make me smarter and take me further. It was like having an expert mechanic work alongside me, and I feel like I’m moving from being an apprentice to being a more capable mechanic.

Last year, I became more intentional about my bucket list items and making sure I don’t sacrifice the things that really matter to me. In a very long conversation with Claude, we discussed about 90 bucket list items I had collected over the years and categorized them into groupings that have now become important concepts of my life. Claude was helpful in identifying patterns in my own bucket list items and in a collaborative process, we identified areas that are meaningful to me. It was helpful to have what seemed like an objective voice on the outside of me, reflecting back some of the things I took for granted or patterns that I’d neglected. It has helped me to be more intentional about relationships and parenting advice in private, and just getting outside to go camp that make me a better person in my work and in my interpersonal relationships. This has helped me to delete some of the things that were noise in my life and identify the things that really mattered. I think that’s made me a more balanced person and people around me have said that I seen more at peace. I could’ve done all this stuff on my own, but it was nice to systematically explore those bucket list items in a back-and-forth way.

I am a professor and I have only had a few students use AI to produce content they claimed as their own. But I have several colleagues who have been very frustrated that students have not taken the time to produce their own work—instead, using AI tools to complete assignments and then move on to the next thing. It is hard for educators who care deeply about developing human beings and coaching students to grow. It feels like it challenges professors’ identity as educators, and then also a great deal of disappointment wells up because they see how students could be so much more but are choosing to shortcut their learning. Again, I’ve only witnessed this a few times myself, but it is disheartening and it feels like it cheapens the effort I put into facilitating an engaging and transformative learning experience for students. 

I never use AI to produce 100% of final written content that goes out to the public that’s supposed to be purely ME. This is especially important for any social media content or content that I put on my website, or emails to people I lead. Authenticity is crucial, and using AI to produce this final content feels like I don’t value my readers because I didn’t do a substantial amount of the work. The care I put into writing something shows how much the reader means to me. That doesn’t mean I don’t use AI for written text that is more operational, but even then, I have tried using AI tools to produce that content as a first draft, and I wind up erasing most of it and writing the rest from scratch.

Gosh, that’s a good question because it has emerged over time. Interestingly, I started out using AI to produce low-stakes content and keeping it more often than not. That was probably a year and a half or two years ago. But in the last year, I feel like my own personal style guide and boundaries have developed through practice. Let me say it this way, every time I write, I am shaping my own AI policy and it tightens in ways based on the context. Every new thing I write for a different audience tests out how AI can assist in that process in ways it can be damaging. This makes me think, I wonder how we can develop this in the people we lead or in students. It’s a metacognitive approach to defining one’s own standards. This requires being aware of one’s own thinking, but if we could do that, that’s a developmental approach that could be quite promising. This gives me a lot to think about.

AI tools are extraordinary at identifying patterns, so I think there’s so much hope when we look at systemic design for AI to assist identifying patterns in areas like healthcare with disease, or climate and weather patterns and drought, or my favorite, and, human brains likely cannot wrap themselves around how complex structures are and their interconnections. AI can help us to better understand the universe and ourselves. I have great hope there.

AI that does not enhance people’s potential should be avoided at all costs. If AI tools create barriers that prevent interpersonal relationships, or if AI tools become a surrogate for deep connection with other people with ourselves, we risk losing the ability to be human. There is a wondrous messiness and developmental opportunity when people create from the very beginning to the very end. The struggle of creating something is how we learn about our craft and about ourselves. So, I struggle at times with how AI can augment or supplement the way we create and sometimes how it could short circuit the benefits of the messy creative process. This applies to science and discoveries, as well as the arts and design—creating a logo or a poem.

So much of AI use is about intent and self-control. As a designer, if I were at anthropic, I might explore ways for AI tools to remind users of ways AI use might reduce the developmental benefits of solving a problem that’s complicated. This is a tricky one though. Because designing a tool that has limits steers people to a certain kind of outcome or use. I don’t love limits put on tools, but I do see some benefit to making explicit the process of use and what someone would get out of that use, or would sacrifice because of use. Perhaps, when people use an anthropic tool, the tool might give users an option for for Bitcoin users is to use the tool to discover during the process of use. It’s almost like there’s a warning label or some kind of contextualization on the use experience before beginning. It would intent. Gee, there’s definitely a theme in my responses, aren’t there?

Oops. I am using voice to text, and that sentence did not render correctly. I will correct it below:

Perhaps, when people use an Anthropic tool, the tool might give users an option for the mode of use before the session starts. Options like 1) discovery:   use the tool to expand thinking and produce insights; 2) operational: use the tool to get something done in a short amount of time. Etc.

I appreciate that Anthropic is doing a study like this. As a researcher myself, it is important to involve participants in the co-creative process. Though participatory design complicates the design process and adds a lot of drag and time, AI has such significant benefits and downsides potentially that it is important for users from many different backgrounds to be involved in that creative process. I understand it’s a balance. But I encourage Anthropic to continue seeking people’s insights.