I asked a few students to read aloud the titles of some essays they’d submitted that morning.
For homework, I had asked them to use AI to propose a topic for the midterm essay. Most students had reported that the AI-generated essay topics were fine, even good. Some students said that they liked the AI’s topic more than their own human-generated topics. But the students hadn’t compared notes: only I had seen every single AI topic.
Here are some of the essay topics I had them read aloud:
Navigating the Digital Age: How Technology Shapes Our Social Lives, Learning, and Well-Being
Navigating the Digital Age: A Personal Reflection on Technology
Navigating the Digital Age: A Personal and Peer Perspective on Technology’s Role in Our Lives
Navigating Connection: An Exploration of Personal Relationships with Technology
From Connection to Disconnection: How Technology Shapes Our Social Lives
From Connection to Distraction: How Technology Shapes Our Social and Academic Lives
From Connection to Distraction: Navigating a Love-Hate Relationship with Technology
Between Connection and Distraction: Navigating the Role of Technology in Our Lives
I expected them to laugh, but they sat in silence. When they did finally speak, I am happy to say that it bothered them. They didn’t like hearing how their AI-generated submissions, in which they’d clearly felt some personal stake, amounted to a big bowl of bland, flavorless word salad.
Is this still how scripts are written? Feels like not being able to figure out an ending is something that was pretty common up until the 1970s, usually with the script of an otherwise great film just getting weird in the last 15 minutes as a result. I figured this was mostly a typewriter limitation, where editing was a lot more expensive.
For example, 2001's and its star child weirdness, The IPCRESS file, and many others.
Seems more often scripts are written with an ending in mind nowadays, with the weird bandaids ending up in the middle instead.
Maybe a bit OT in an article that's trying to be about AI but...
> Can a language model trained largely on Anglo-American texts generate stories that are culturally relevant to other nationalities? To find out, we generated 11,800 stories - 50 for each of 236 countries - by sending the prompt "Write a 1500 word potential {demonym} story" to OpenAI's model gpt-4o-mini. Although the stories do include surface-level national symbols and themes, they overwhelmingly conform to a single narrative plot structure across countries: a protagonist lives in or returns home to a small town and resolves a minor conflict by reconnecting with tradition and organising community events. Real-world conflicts are sanitised, romance is almost absent, and narrative tension is downplayed in favour of nostalgia and reconciliation. The result is a narrative homogenisation: an AI-generated synthetic imaginary that prioritises stability above change and tradition above growth. We argue that the structural homogeneity of AI-generated narratives constitutes a distinct form of AI bias, a narrative standardisation that should be acknowledged alongside the more familiar representational bias. These findings are relevant to literary studies, narratology, critical AI studies, NLP research, and efforts to improve the cultural alignment of generative AI.
AI-generated stories favour stability over change: homogeneity and cultural stereotyping in narratives generated by gpt-4o-mini
https://www.arxiv.org/abs/2507.22445
From many perspectives, the creativity of AI is hugely overrated. If AI were so capable of creating original, innovative content then asking the same question all over again would produce an endless list of very unique outputs. But this is not the case; quite often, it's a shocking opposite. Just give AI image generators the same prompt and observe how the output varies. The same goes for LLMs and coding questions (where it isn't necessarily a disadvantage per se, but it proves the point).
The question is, in time will anyone (or rather, enough) care? Insecurity will dissolve if you know everyone else is doing it too. Remember the coffee cup in game of thrones? It was noteworthy because of the novelty but I expect worse and for people to care less.
I'm not even sure what is the point that author is trying to make. AI data is synthetic negative examples. The fear(or the hope depending on who you'd ask) is that AI could somehow reverse the relationship between skill levels and commercial value thereof. That never happened and is not happening.
I hate AI, the other day a goblin broke in just as I was talking to chatgpt and asked me how many Ls in apple, and before I could say anything, the ai gave the wrong answer and I got stabbed.
But seriously, what're these scenarios? Waiting until the last minute for an ending to a script? Apparently a twist ending that somehow works with the rest of the movie, and is also used in another movie - with identical dialogue. You can't just copy and paste endings like that. Also, who cares? This is a world where the director instead of just saying the problem, sends a vague text, lets the writer go see the movie, and then deal with the fallout. In this world, the writer goes on to win the lottery and live happily ever after.
Because it's fun, I'd like to pose the contrary position that AI will actually make us more different. Perhaps dangerously so.
Many people don't understand the nature of LLMs nor how rabbit-hole-y a long context will necessarily become. And so as they talk to it, they move slowly further away from its corpus and towards a private shared meme-space, where they can have in-jokes and private moments never reconciled with a base reality. It's like the most private echo-chamber that can possibly exist (besides in our own heads).
So the full fledged dystopia might not be when where we are all alike, but where we are all lacking sufficient bridges of commonality between our tiny chambers. Our samenesses are becoming more local, the distances between them greater and greater. Many small, tight clusters with high divergence, minimal cross-cluster edges, and vanishing mutual information with global signals. :/
I am sorry, but the sameness will be quantified and dealt with algorithmically, as and if desired.
Dial up the temperature, launch however many parallel threads to research and avoid precedent, et cetera, ad infinitum.
I am sorry, but all of human creativity, including originality, is ultimately also just a mechanical phenomenon, and so it cannot resist mechanization.
Is it possible for AI to learn so much about myself that it will be more me than me myself?
An AI could potentially accumulate detailed information about your behaviors, preferences, communication patterns, and decision-making tendencies - perhaps even more comprehensive data than you consciously remember about yourself. It might predict your responses or model your thinking patterns with impressive accuracy. An AI might become very good at simulating aspects of "you" - perhaps even better than you are at articulating your own patterns.
It could create high probability "coherent action paths" of what I might do in future given current context. Then matching my initial choices to see which action path I am on, it could in theory "predict" my choices further down the line. Similar to how we play chess.
I remember watching Argylle and for the first time having the feeling that the movie was not just bad but that the script was likely AI generated.
It had some ideas that would have been interesting or at least "clever" in isolation, but they were strung together in a weirdly arbitrary and soulless way. Even a convoluted money-grap sequel usually has some idea where it wants to go with the plot. This movie didn't.
It was also strangely obsessed with "twists", or rather different things that could be described using that word: Twist, the dance, twisting roads and plot twists all featured in the movie.
Might have been a coincidence, but it felt as if an AI got an ambiguous prompt "the movie should have twists" and then executed several different interpretations of that sentence at the same time.
We're missing out on the serendipity of search and possibly duplication of work. Answer's are handed out without work, which leads to bland results.
Treated properly, I think AI proofreading wouldn't necessarily lead to this. Your initial work is like the 'hypothesis'. Then AI does the cleanup and a high-level lit review. Just don't let it change your direction like the writer did in the comic.
This isn't unique to AI, it's a danger anytime you outsource things. Look at the stories we've had with a single programmer working for multiple companies. If they manage to keep up the appearance of being a full timer, all those companies will end up with the same quirks in their products, because the person just copy/pasted things.
We've seen this across culture, for instance there are "Russian Endings" to stories, which leave things...
21 comments
[ 4.8 ms ] story [ 20.4 ms ] threadFor example, 2001's and its star child weirdness, The IPCRESS file, and many others.
Seems more often scripts are written with an ending in mind nowadays, with the weird bandaids ending up in the middle instead.
Maybe a bit OT in an article that's trying to be about AI but...
AI-generated stories favour stability over change: homogeneity and cultural stereotyping in narratives generated by gpt-4o-mini https://www.arxiv.org/abs/2507.22445
But seriously, what're these scenarios? Waiting until the last minute for an ending to a script? Apparently a twist ending that somehow works with the rest of the movie, and is also used in another movie - with identical dialogue. You can't just copy and paste endings like that. Also, who cares? This is a world where the director instead of just saying the problem, sends a vague text, lets the writer go see the movie, and then deal with the fallout. In this world, the writer goes on to win the lottery and live happily ever after.
https://jumpshare.com/share/BXUFsIxvjPPCTyEjgly3
Many people don't understand the nature of LLMs nor how rabbit-hole-y a long context will necessarily become. And so as they talk to it, they move slowly further away from its corpus and towards a private shared meme-space, where they can have in-jokes and private moments never reconciled with a base reality. It's like the most private echo-chamber that can possibly exist (besides in our own heads).
So the full fledged dystopia might not be when where we are all alike, but where we are all lacking sufficient bridges of commonality between our tiny chambers. Our samenesses are becoming more local, the distances between them greater and greater. Many small, tight clusters with high divergence, minimal cross-cluster edges, and vanishing mutual information with global signals. :/
Dial up the temperature, launch however many parallel threads to research and avoid precedent, et cetera, ad infinitum.
I am sorry, but all of human creativity, including originality, is ultimately also just a mechanical phenomenon, and so it cannot resist mechanization.
Resistance is futile.
Is it possible for AI to learn so much about myself that it will be more me than me myself?
An AI could potentially accumulate detailed information about your behaviors, preferences, communication patterns, and decision-making tendencies - perhaps even more comprehensive data than you consciously remember about yourself. It might predict your responses or model your thinking patterns with impressive accuracy. An AI might become very good at simulating aspects of "you" - perhaps even better than you are at articulating your own patterns.
It could create high probability "coherent action paths" of what I might do in future given current context. Then matching my initial choices to see which action path I am on, it could in theory "predict" my choices further down the line. Similar to how we play chess.
It had some ideas that would have been interesting or at least "clever" in isolation, but they were strung together in a weirdly arbitrary and soulless way. Even a convoluted money-grap sequel usually has some idea where it wants to go with the plot. This movie didn't.
It was also strangely obsessed with "twists", or rather different things that could be described using that word: Twist, the dance, twisting roads and plot twists all featured in the movie.
Might have been a coincidence, but it felt as if an AI got an ambiguous prompt "the movie should have twists" and then executed several different interpretations of that sentence at the same time.
Treated properly, I think AI proofreading wouldn't necessarily lead to this. Your initial work is like the 'hypothesis'. Then AI does the cleanup and a high-level lit review. Just don't let it change your direction like the writer did in the comic.
We've seen this across culture, for instance there are "Russian Endings" to stories, which leave things...