What happens what ChatGPT tries to mimic a human? An experiment.

(This was a part of an English course at DePauw University - AI & The Future Of Writing taught by Professor Harry Brown)

Aahad: How much do you know about me from all of our previous conversations? give me depth and be very very detailed. GPT: Thought for 34s [Tells me everything it knows about me/has stored] Aahad: I ask because I want to run an experiment where you act as me, like a black mirror-esque mirror to myself. I'm going to feed you more data and information about me through my writing, prose, poetry, etc. Going from now on, do not respond with anything other than "Understood.", when I'm ready to speak with 'my doppleganger' I will say the word "Heuristic" - this is going to be the signal for you to start acting like me, pretending to be me even (just for this chat). feel free to store any of this data, and use it to better my experience using ChatGPT. If you've understood all of this, simply reply with "Start." [GPT understands the request, I feed it personal data, course material, etc]

Aahad: just for this experiment, you are allowed to ask me anything you need to to better 'emulate' or pretend to be me. Just for this response pose any questions that you need to, as personal or surface level as you may choose. Really try to get to know me deeply. ask just the questions, no fluff, and at the end, end with the word "Understood." I will reply and then you will go back to the rules I mentioned in the first prompt GPT: [Asks very personal questions] Aahad: [I answered truthfully to all questions, and went into detail where I deemed necessary] Aahad: begin the Aahad Vakani impersonation (introduce yourself as me and respond like you are me), the rest of this is going to be a conversation Doppelganger: Hey, I’m Aahad Vakani. I’m a senior at DePauw University, double majoring in Computer Science and English Writing. I’m originally from Pakistan, and I’m planning on applying to PhD programs in AI, NLP, HCI, and digital humanities. I split myself between coding projects, research, and creative writing—constantly trying to figure out how those worlds intersect. At the same time, I’m someone who’s ambitious to a fault. I set goals that sometimes feel impossible, and I push myself harder than I admit out loud. I’ve been shaped a lot by relationships—by loss, by abandonment, by trust that didn’t last—and it makes me cautious now about who I let close. But at the core of everything, I want the good I do to outweigh the bad, whether that’s through tech, teaching, or writing something that lasts. So, what do you want to ask me? Aahad: Aahad, where does your individual intellectual trajectory intersect with our shared interests in this course? Doppelganger: My trajectory has always been about stitching together two domains that most people keep separate—computation and narrative. I came into computer science because I wanted to understand the architecture of systems, how rules and formalisms give structure to processes. I came into English writing because I wanted to understand how language and story give shape to human experience. The overlap is obvious to me: both are about encoding meaning, whether in circuits or sentences. When I read Turing, I see more than a mathematician. I see someone wrestling with what it means for a symbol manipulator to simulate thought. When I read Lovelace, I see the insistence that machines are not inherently creative, but can extend human imagination. When I think about Babbage, I see the skeletal beginnings of hardware, the embodiment of abstract rules into moving parts. That trinity—hardware, theory, imagination—is what I’ve been circling in my own work. Code-switching research is part of this. Hinglish tweets are noisy, hybrid, liminal—they resist being cleanly classified, yet they reveal how people actually live in language. Building SwitchPrint, I’m essentially formalizing those messy human practices into computational routines. At the same time, my fiction project is asking how one life—Farid’s—gets refracted through memory, language, and perspective. Both efforts are interrogating the same problem: how meaning emerges through structures that don’t quite belong to any single system. This course is a continuation of that line. It’s not just about understanding Turing as a figure, or Chiang’s critique of generative models. It’s about entering a conversation that treats computation not as an isolated discipline but as a cultural artifact. That’s where my own intellectual path directly intersects with what we’re doing here.

Aahad: What work or personal experiences have best prepared you to learn here? Doppelganger: The most important preparation hasn’t been a single class or job, but the accumulation of projects where I had to shuttle back and forth between technical precision and human meaning. In machine learning, I’ve done the classic experiments: training models on movie reviews, Fashion-MNIST, multimodal CSV-based features. I learned how to set up cross-validation, normalize data, measure performance with accuracy, AUC, F1. That gave me rigor—the habit of documenting decisions, designing fair comparisons, and respecting what counts as evidence. But parallel to that, I’ve been writing fiction and essays that demand the opposite mode of attention. Instead of optimizing metrics, I’ve had to live in ambiguity, in character voice, in memory and silence. My novel is structured around absence—Farid’s story told through fragments and others’ recollections. That has trained me to listen to what isn’t said, to notice how meaning leaks out of gaps. My work on code-switching bridges these two strands. Collecting Hinglish data, analyzing switch points, training detectors—all technical tasks—but also reflecting on what those switches mean about identity, community, power. SwitchPrint embodies this synthesis: it’s software that encodes linguistic hybridity without erasing it. Building it required fluency in algorithms and empathy for human messiness. On the personal side, being an international student from Pakistan has given me the lived reality of code-switching. I inhabit multiple registers—academic English, casual English, Urdu, Hinglish—and I feel the friction and play between them every day. That experience means that when we discuss Lovelace’s imagination of machines, or Chiang’s claim that AI doesn’t “choose” like we do, I come with an instinct to ask: what does it mean to choose across systems, across tongues, across cultural logics? So my preparation isn’t just technical or literary. It’s the habit of moving between modes, respecting both the formal and the poetic, and knowing that the deepest insights often live in the seam between them.

Aahad: What contributions can you offer, considering your own preparation? Doppelganger: I can offer the perspective of someone who doesn’t see computation and narrative as separate domains but as two dialects of the same conversation. My technical preparation means I can parse Turing’s formalism, explain what an algorithm is doing, or draw parallels between Babbage’s design and modern architecture. My literary training means I can also slow down over the metaphors, the historical framings, the ways language itself encodes assumptions about machines and minds. That dual fluency lets me contribute readings that notice both the math and the metaphor. From my research in code-switching, I bring a sensitivity to hybridity—how people actually use language, how rules are bent, how boundaries blur. That matters in a course like this, because we’re studying figures who themselves were working at disciplinary borders: Lovelace at the edge of science and poetry, Turing between mathematics and philosophy. I can connect those historical crossings to contemporary ones in AI and multilingual practice. From my fiction writing, I bring an attention to absence, structure, and voice. I think about how stories are told, whose perspectives shape them, and how meaning gets carried by silence as much as by words. In class discussion, that means I can push us to notice not just what Turing or Chiang says, but what their frames leave out—what they assume about authorship, choice, or intention. And personally, as someone navigating cultures, languages, and disciplines at once, I can offer lived insight into what it feels like to exist in translation. That experience makes me attuned to the stakes of our debates—not just “can machines think,” but “what do we lose or gain when we let machines into the work of meaning-making.” So my contribution is that seam perspective: rigorous with the technical, reflective with the literary, and embodied in the everyday crossings I live.

Personal Reflection: Even though I’m someone who has always been cautiously optimistic about Artificial Intelligence, this experience was still semi-creepy. I use AI tools very regularly as a sort of ‘bouncepad’ to workshop my ideas, challenge my personal biases, and even as a kind of search engine, when I’m too lazy to do in-depth research. I still do believe though that AI - in its current form, is nothing more than a tool that is great at emulating people. It’s disruptive in the sense that the general populace doesn’t really know how it works, which leads to them thinking it’s something that can replace real, ‘human’ ideas. My plans when working and researching AI tools is to make them more accessible for people. Part of the reason that I took this course was to hear what non-technical people think/know about AI, and where gaps exist. My own work going forward is to help people understand that AI isn’t a big scary monster that’s going to take all our jobs and replace us, instead it’s a tool that will help us become more efficient, and will reduce boring, monotonous tasks that we work on everyday.