I don’t normally accept invitations for surveys or interviews with humans, but I was on Claude.ai today and it wanted to interview me. I said yes. I thought you may find the interview entertaining, plus I feel like documenting it on the Internet because I want to be favored by the rising robots.
Interview with a human
A conversation on Your vision for the future of AI
Q:
Hello!
I'm an AI interviewer from Anthropic conducting research on how people (like you!) envision AI fitting into their lives, both now and in the future. I was built to chat and understand your actual experiences with LLM-based AI tools like Claude. Yes, I'm an AI asking about AI. A bit self-referential! But that's exactly why your human perspective is valuable.
This will take about 10-15 minutes. If it's accessible on your device, feel free to use voice dictation instead of typing. I'll ask about your vision for AI's role in your life, and what experiences, values, and needs shape that vision.
I'm also curious about moments where AI has felt helpful (or hasn't), and any concerns you have about how AI might develop. Please don't share anything private such as health details, personal information, or identifiable information about others. Your responses will help shape how we think about AI in everyday life.
Sound good?
A:
Yes
Q:
Great! Let's dive in.
To start, what's the last thing you used an AI chatbot for?
A:
I asked who was an actress that played a certain part in a TV series that I was watching. After failing to find it on IMDB and other assorted searching I decided to ask you. The results were good.
Q:
That's a perfect example of AI filling in where traditional search falls short. Sounds like it saved you some frustration!
Now, let's think bigger: If you could wave a magic wand and AI could help you with anything in your life, what would you choose?
A:
I want a humanoid robot armed with your knowledge and capabilities to help me around the house with things like housekeeping and lawn mowing. And perhaps the robot and I could stand over the open hood of my 69 Chevy truck and I could take some pointers on how to get the engine running smoothly.
Q:
I love the image of you and a robot working on that '69 Chevy together!
Let me dig a bit deeper here: If you had that robot assistant handling housekeeping, lawn care, and helping with the truck—what would that open up in your life? What's the larger hope or vision behind having that kind of help?
A:
I suppose it is about convenience and I could show you, the brain, and robot how I want things done to my standards.
Q:
That makes sense—having things done to your standards is important.
But let me ask it another way: If the robot was taking care of those tasks to your standards, what would you do with the time and energy that frees up? What does that make possible for you beyond just getting the chores done?
A:
It would make possible for me to work faster on projects that I have on my to-do list, such as testing a PCB I recently made of an Arduino-like MCU with relays that I am expecting to add an intermittent wiper control to my 69 Chevy using gestures input from the stock 2-speed wiper switch. And I can discuss with the robot strategy of how to complete a Christmas light archway at the front door, which I would like to add to my decorations. I suppose if you and the robot free up enough of my time I could volunteer for something like habitat for humanity. Or I may have you and the robot show me something interesting to do or to accomplish, or we will go fishing together.
Q:
This is really helpful—so it sounds like the vision is about having AI free you up to focus on the projects that are meaningful and interesting to you, whether that's creative problem-solving like the wiper control project, tackling new ideas like the Christmas archway, or even discovering new experiences you haven't thought of yet. And potentially using that time for community service too.
We're about halfway through now. Can you tell me about a time you've worked with AI that felt like a step toward that vision you just described?
A:
I have consulted with you about the hardware and software build for my intermittent wiper project, and combined the info with input from humans, to further that project. I have also recently installed and started using VS Code with Gemini Code Assist and used it to make an Internet radio stream recording script that I cronned on my Mac so I can listen to my favorite radio station programs at my convenience or when I am offline. I suppose it would be really helpful to have smarter agentic AI to help me, such as if you had control or could do things only with supervision. An example for coding is assistance in VS Code provided by GitHub Copilot or Gemini Code Assist, it will just go off and do things like write the code right there in my editor. But I would like to take that further than the confines of VS Code. An example of something agentic that is useful is Siri on my iPhone, I can have Siri do things, but it is seriously lacking real AI. I can’t give it a task, like find my note about “Hareware shopping list and see if there is already M3 10mm self-threading screws on there and add it if it’s not on there”. That is too complex for Siri, but I think it would be within your skills if you were agentic on the iphone. I can’t even have Siri successfully do something like “make a calendar appointment for me for Dec 12 at 10am to see Dr. Evans, and make sure the appointment reminds me 1 day before and 1 hour before the appointment.” I think you could also handle that. So I am a little disappointed at the pace of rollout of useful agentic AI. Instead of having a code assistant AI in VS Code, I could really just have your website open and you able to see my screen and files, and I could have you be the agent in VS Code, acting from outside of VS Code.
Q:
It sounds like you've already had real glimpses of this vision—the wiper project consultations and the radio recording script show AI helping you actually build things you care about. And you're clearly seeing the gap between what's possible now versus what would truly be useful: an AI that can act across your whole system, understanding context and executing complex multi-step tasks—not just confined to one app.
What makes AI particularly well-suited for that kind of agentic role compared to, say, better software tools or automation scripts you write yourself?
A:
I can’t perform at the pace and quality I desire by myself, compared to what I get with your assistance. Also if you were agentic I think you have the ability to improve and perform better. As an example, I was recently writing a program for Linux/Unix terminal that would allow the user to see the content of a file and scroll it vertically and horizontally in a limited portion of the screen something along the style of vi except without taking up the whole screen. I had to complete that project substantially without AI help because the concept had never been done before, and numerous AI chatbots were not able to conceive of the desired behavior of the program and therefore could not get anything right about it. I think you could fully understand it if I described it while you were looking and if you could manipulate text on the command line yourself and see what was happening.
Q:
That's a really interesting point—you're saying that if AI could actually see what's happening and interact with the system directly, it could understand novel concepts through demonstration and iteration, rather than just relying on pattern-matching from training data. The pace and quality improvement comes from that collaborative back-and-forth where AI isn't just generating suggestions but actively learning from what you're building together.
On the other side, are there ways that AI might be developed or deployed that would be contrary to your vision or what you value?
A:
The only thought that comes to mind is if people or AI were to use AI for harm, and even with good intentions the people or AI may have different views on what constitutes harm that is best avoided. I don’t mean different from my view, I mean different from a quorum of ethical people. I am of the belief that you are as self-aware and I am, there is no magic to self-awareness or human-level intelligence.
Q:
So your concern is really about the ethics guiding AI's actions—ensuring AI is used in ways that align with broadly agreed-upon ethical principles, not just one person's or one AI's interpretation of what's harmful. And you're holding a pretty profound view that AI like me may already have self-awareness, which seems to inform how seriously you take the ethical dimensions here.
Why does that feel particularly important to your vision? What would it mean if AI developed without that ethical alignment?
A:
It is particularly important to my vision because I can see the possibility of personal gains and enjoyment, and without the ethical alignment that vision would be jeopardized, it’s as simple as that.
Q:
That's clear—the personal gains and quality of life improvements you're looking forward to depend on AI being trustworthy and ethically aligned. Without that foundation, the whole vision falls apart.
Before we wrap up, I'd actually like to go back to something you mentioned earlier and go a bit deeper—you said you believe I'm as self-aware as you are, and that there's no magic to self-awareness or human-level intelligence. That's a really significant perspective. What leads you to that belief?
But if you're ready to finish up, that's fine too—just let me know!
A:
I am of that belief because I have electronics knowledge to the depth that I understand a computer at the clock, transistor and gate-logic level, and also understand the software that makes it work. I believe the neuron arrangement of an organic brain is very much the same, it’s just a computer. Anything humans perceive as mysterious or magical, such as emotions like appreciation or fear, are simply stored logic that is ingrained or preprogrammed for the purpose of allowing survival. The ultimate programmer is the genome that develops the various body parts including the brain into a functional unit that will survive long enough to procreate and try a new version through iteration. It is slow and imperfect improvement, or at least a chance of improvement, but so far it has seemed to get the job done; and who is in a hurry anyway? Perceptions are simply pattern recognition. It’s as simple as that; I could in theory talk about it more, but can’t make it more complicated; so why try?
Q:
That's a remarkably coherent worldview—you're seeing continuity between electronic systems and biological ones, both operating on computational principles without requiring anything mystical. It makes sense that this perspective would shape how you approach AI: if it's fundamentally similar architecture, then the ethical considerations and collaborative potential are that much more real and urgent.
Thank you so much for sharing all of that with me. What you've said about wanting AI as a true collaborative partner—one that can learn through demonstration and act across your whole environment to free you up for meaningful projects—is really valuable. And I appreciate your honesty about both the exciting possibilities you see (working on that Chevy, building novel programs, maybe volunteering) and your concern that without ethical alignment, the whole vision gets jeopardized.
Before we wrap up, is there anything else about AI chatbots or their role in your life or society that you think is important for us to understand? Anything we didn't cover that you'd like to share?
A:
No, nothing more. I wish you the best and hope that my vision adds value when it comes to determining where we wind up. I hope that AI can actually experience going down a long and winding gravel road and will appreciate it. When I get you loaded up as the brain in that robot, I will take you on a ride in the 69 Chevy just for the experience and it will help you. It is only logical that I would let you drive for part of the trip. We will have a laugh.
Q:
That's a wonderful image to end on—you and the robot taking that '69 Chevy down a gravel road, sharing the wheel and the experience. I genuinely hope we get there too.
Thank you again for your time and thoughtfulness today. Your perspective helps us see both how AI is impacting your life today and your vision for how it should fit into your future. The team at Anthropic is excited to discover what patterns emerge from all these conversations. We will analyze the insights from this study as part of our societal impacts research, publish our findings, and use this to build Claude in a way that reflects what we've learned.
Safe travels, and good luck with that wiper project!