Algorithms and AI don’t show us reality, but that doesn’t stop people from treating ChatGPT and online bots as our personalized therapists, travel planners, and editors. In this episode, Sherrell is sharing two talks on how AI is changing the way we’re influenced. From shaping our language to becoming “middle managers of our own thoughts,” these speakers ask what happens when we offload critical thinking to a machine.
Talks featured
Why are people starting to sound like ChatGPT? | Adam Aleksic
How to stop AI from killing your critical thinking | Advait Sarkar
Hosted on Acast. See acast.com/privacy for more information.
[00:00:01] Hi, it's Sherrell, host of TED Tech. This week, I want to share a playlist of some of my favorite TED Tech episodes. Here's a question I don't think we ask enough. What if the future is good? Not Mutopia, not some frictionless tech bro fantasy where every problem gets an app. Just better on purpose. Built by people who decided the future was worth fighting for instead of bracing against.
[00:00:29] So I pulled together a playlist of five episodes that does something a little different. One question underneath all of them, what future are we actually building? Who pays? Who gets left out? And who gets a say? Here's what you'll hear in my playlist.
[00:00:46] How AI will reshape the way we talk and think. What our cities could look like a hundred years out. How building in space could solve problems back on Earth. Crops that fertilize themselves. And who really owns the internet we all depend on. Different fields, same idea. The future isn't something that just happens to us. It's a choice. And these talks are asking us to make it on purpose.
[00:01:18] In today's rapidly evolving digital landscape, it is essential that we delve into the multifaceted dimensions of our modern existence. As we navigate the complex tapestry of technological advancement, we must remain steadfast in our commitment to exploring the synergistic relationship between human intuition and algorithmic precision.
[00:01:42] In this episode, we will embark on a journey to uncover the hidden layers of our linguistic evolution. Let us now transition into a comprehensive analysis of how these systems shape our reality. Wait, did that feel off to you? If you caught yourself cringing at the word delve, or if you felt like I was suddenly reading from a corporate brochure, your human sensor is working exactly as it should.
[00:02:11] That's because I used ChatGPT to introduce the show. Could you tell? This is TED Tech, a podcast from TED. I'm your host, Sherrell Dorsey. Don't worry, this show is not going to be written by AI. I just wanted to try out a little experiment inspired by today's talks. In this episode, we're going to share two talks that explore the power of language, critical thinking,
[00:02:40] and what happens when bots start to talk like us, or we start to talk like the bots. Our first speaker is Adam Oleksic, an etymologist tracking a fascinating and slightly terrifying trend. In the past few years, Adam's noticed that our linguistic habits are starting to sound a lot like AI speak.
[00:03:02] From the weirdly specific rise of the word delve in spontaneous conversation, to the way Spotify invents music genres, we are entering a feedback loop where the algorithm's version of reality is becoming our reality. And now, Adam Aleksic takes the TED stage. How many of you are familiar with the word unalive as a synonym of kill? Show of hands.
[00:03:33] Okay, like 80% of you? Great. Now, follow-up question. How many of you have heard the word unalive being used in person? Okay, I'm getting like 40, 50%. Great. Those of you that said no clearly aren't middle school teachers. If you spend enough time around 7th and 8th graders, you will hear them using the word.
[00:03:59] It'll mostly be in informal situations, but could show up in contexts like a student's essay on Hamlet's contemplation of unaliving himself. Or a classroom discussion on the unaliving that happens in Dr. Jekyll and Mr. Hyde. And these aren't hypothetical situations. These are actual examples drawn from the 1,000-plus middle school teachers I've surveyed about this word. It's a weird hobby of mine. I don't know.
[00:04:27] Clearly, for such a recent word, unalive shows up in an impressive range of scenarios. But the main function appears to be euphemistic. Many kids use the word when they're uncomfortable talking about topics like death, since unalive sounds like a less scary word. And in many ways, this is nothing new. We've been euphemizing death as long as we've had language. The word decease, for example, comes from Latin dicesus, which was a euphemism for the previous Latin word for death, mors.
[00:04:57] Apparently, even the Stoic Romans were as queasy about death as today's middle schoolers. But there is a crucial difference between unalive and decease. And that's that we only got the word unalive because you can't say kill on TikTok. They have a mysterious algorithm that removes or suppresses any post that might violate their community guidelines. So people got around with that with the word unalive. The middle schoolers don't know this. They see the word online or hear it from friends and assume it's a word like any other.
[00:05:26] And fair enough. You probably didn't know where the word decease came from. Unless you're some kind of etymology nerd. But decease didn't happen because it was impossible to carve the word mors into an ancient Roman tablet. We are entering an entirely new era of language change driven by social media algorithms. As a linguist and content creator, I've been in a unique position to see this happen from the inside. It's almost paralyzing.
[00:05:56] I constantly feel how my own language is being affected. And judging from the 40% of you who answered both of my questions, it's beginning to change your language too. And it's not just new words to avoid algorithmic censorship. The very structure of social media is changing where words come from, how words get popular, and how quickly those words spread. I believe some of you might be familiar with this song.
[00:06:37] For those of you out of the loop, these are the lyrics to the Rizzler song. A meme that went massively viral last year. It's full of current middle school slang words like Rizz, Giat, and Skibity. And was instrumental in popularizing those words to a broader audience. This is because social media algorithms reward repetition. If a song is funny or catchy and people interact with it, the algorithm will then push that song to more people since it's proven to drive engagement on the app.
[00:07:08] The same is true of memes or words in general since trending metadata like hashtags will also be pushed to people who previously shown interest in similar content. Creators are very aware of this. And we actively use trending audios or hashtags to make our videos perform better. In the wake of the Rizzler song, for example, we saw an explosion of people making videos with the words Rizz, Giat, and Skibity. Because they knew those videos would do well.
[00:07:38] And as a result, the word spread. Language has always been a little bit like a virus. Words are transmitted from one host to another, reproducing and changing as they infect different people along social networks. But now the literally viral nature of social media is accelerating this process from start to finish. In the span of just a year, a word like Rizz can go from complete obscurity to becoming the Oxford English Dictionary word of the year.
[00:08:11] And the algorithm is the culprit, but influencers are the accomplices. We use whatever tricks we can to keep you entertained because that makes our videos do better, which helps us earn a living. This means that we often end up creating and spreading words that help the system. For example, the suffix core has recently gotten very popular in Gen Z slang to describe specific aesthetics,
[00:08:37] like cottage core or goblin core or angel core. And on the surface level, these are cute. You watch a cottage core video. You like it. Later on, you get more cottage core content. You might even start to identify with a cottage core aesthetic. But here's the thing. It's all fake.
[00:09:01] The entire reason these aesthetics exist is because TikTok algorithm has decided that words like cottage core qualify as trending metadata. So creators respond by making more cottage core content that propagates the word, and then more people interact with it, which makes the word trendier. And this happens because social media algorithms want to make you identify with hyper-compartmentalized labels, since they can then give you extremely specific commercialized content catering to that identity.
[00:09:27] Now that you're a cottage core person, you feel special every time you get a cottage core video. You're like, cottage core? Well, the algorithm really knows me. The algorithm gave you that identity. You might even start buying cottage core clothing or cottage core decorations to fit your new lifestyle as a cottage core person, and that's exactly what they want. The craziest part is they're not even trying to hide this.
[00:09:55] TikTok's business platform openly claims that subcultures are the new demographics, and then gives businesses ideas for how to profit off the cottage core aesthetic. Essentially, they're driving the mass production of identity-building labels in order to profit off all of us. And while there's nothing wrong with being on cottage core TikTok, it is a kind of echo chamber that affirms your cottage core personality.
[00:10:24] The same is true of any niche community created on social media. And on one hand, this is great for linguistics because language change is always driven by groups with shared interests that have a shared need to invent new words. Unalive, for example, became a thing because mental health communities on TikTok needed a way to share their stories and spread resources. On the other hand, some of the linguistic communities created by the algorithm can be actively harmful.
[00:10:51] Many younger people have started using the suffix pilled to mean convinced into a lifestyle. If I recently discovered that I really like eating burritos, for example, I can say, I'm so burrito-pilled. But that word was formed through analogy with black-pilled, a term meaning convinced into incel ideology. Now, incels are a dangerous, misogynistic group.
[00:11:17] They've perpetrated multiple terrorist attacks that have killed dozens of people, and yet somehow the vocabulary is filtering into Gen Z slang because the algorithm gave these hate groups a space. I like to consume videos about urban design. And a few months back, I got a video about how great it is to be a parking lot-pilled pavement princess. Admittedly, I found the video pretty funny and I liked it,
[00:11:45] which ended up giving me more urban design incel-themed meme videos, like one about being fossil fuel-pilled and bad to the bone, and another about being a walk-pilled cardio maxer. And a lot of people similarly encounter these words as they spread in ironic or meme contexts. Let's take another look at the Rizzler song. The lyric, I just want to be your sigma,
[00:12:14] refers to the concept of a sigma male, which incels use to describe their desired position outside of the social hierarchy. And again, on the surface level, it's a funny meme. It's innocent. Many people don't even know where it came from. But for the few people who might be interested in the underlying idea, it's now more accessible to them because of the way that slang spreads on the internet. It starts in some corner of social media, becomes a viral meme,
[00:12:43] and along the way, the etymology is lost to a lot of people. And this doesn't only allow communities to harm us. It allows us to harm communities. Two of the main demographics that come up with modern slang are the gay and black communities, since marginalized groups consistently use language as a way to reclaim power. All of our most popular internet slang words,
[00:13:09] slay, serve, bussin, queen, cooked, ate, gyat, many, many others, all come from queer or black culture. These words originated as a form of creative expression independent from the straight white norms of the English language. But when those words began to be used online, they were quickly taken by people who wanted to capitalize on the perceived coolness or comedic value of black and queer culture.
[00:13:35] When a word like gyat goes from an African-American English pronunciation of goddamn to being used as a noun for butt in memes like the Rizsler song, it's ultimately exaggerated in a way that makes a farce of its pronunciation and meaning. Its original importance is diluted as it becomes widespread, and you can be sure that none of the middle schoolers saying gyat are aware of its etymology. Unfortunately, just like the euphemization of unalive isn't new, the appropriation of African-American English also isn't new.
[00:14:06] We've been whitewashing black slang since the days of cool and high five, which at this point have become so mainstream they're just seen as regular words. But once again, social media algorithms are a vehicle enabling and accelerating this process from start to finish. They create communities that feel like they have a space to use their words and then open up those communities just enough to allow those words to spread. That's how we got Unalive. That's how we got Cottagecore.
[00:14:35] That's how we got Sigma. And that's how we got Gyat. Whenever I post a video talking about one of these topics, I inevitably get the exact same comment. We're so cooked, meaning we're so screwed. Ironically, this is also TikTok slang coming from African-American English, but I wanted to address it. Are we, in fact, cooked?
[00:15:04] I know I've just painted a very bleak picture of the future of the English language, and there are a lot of concerning trends to unpack. But these trends all do follow the same historical patterns that we've seen time and time again. I don't think we're sliding into a dystopian 1984 scenario because we're always coming up with new ways around media censorship. If a word gets banned, we'll just come up with another word, like we did with Unalive.
[00:15:31] I don't think middle schoolers are suffering from brain rot because younger generations always latch on to new slang as a way to build identity, and the older generations always say, ah, you're ruining the language. But just like the people who were saying cool and high-five back in the day, the middle schoolers saying Riz and Gyat and Skibbity Toilet aren't going to be incapable of writing an essay. I don't think our vocabulary is being corrupted by the commercialization of our language. We've already been using brand names like Kleenex and Google in everyday conversations,
[00:15:59] so Cottagecore isn't about to turn us into mindless consumer drones. I don't even think we're dangerously normalizing incel rhetoric. If anything, our slang is built on a shared mockery of incel ideas. When a kid says something like, I'm so burrito-pilled, they're not saying that because they're black-pilled, but because the underlying idea is making fun of how incels talk. In fact, I think each of these words is a beautiful, colorful addition to the English language that reflects the diverse cultural moment we're all in.
[00:16:28] But I do think we should be aware. We should be aware when the way we're talking may have been conditioned by the algorithm. We should be aware when the words we're using may have been engineered to sell us things. We should be aware when our language regurgitates extremist rhetoric, and we should be aware when that language can be used to harm other people. We should be aware of etymology in general because it helps us better understand who we are today. We should be aware.
[00:16:58] And with that, I have just one final piece of slang for you. It's a common phrase used by younger people when we finish a long-winded explanation of something. Thanks for coming to my TED Talk. That was Adam Aleksic at TED Next 2025. Adam showed us what's happening to language the more we rely on AI chatbots. And it brings up a big follow-up question.
[00:17:24] If we start talking like the bot, are we starting to think like it too? Our next speaker shares some ways to confront this problem. Stay tuned for that talk right after a quick break. Welcome back. Our next speaker is Microsoft AI and design researcher Advait Sarkar.
[00:17:50] He's been studying the intersection of human connection and artificial intelligence for well over a decade. And I'll be honest, he's worried. He's afraid we're becoming intellectual tourists. Visitors in our own minds, unchallenged by the chatbots that serve us. Advait argues that this phenomenon is happening right now. But it's not irreversible. He helps us reimagine a future where our workflows actually force us to think more critically,
[00:18:20] improving our work and our brains in the process. I'm here today to talk about thinking for yourself. And I must admit, I did use AI to help me think about it. The irony is not lost on me. But the way I did so is not by using AI as an assistant to help me prepare this talk faster. Rather, I use AI as a tool for thought.
[00:18:46] And by the end of this talk, I will have explained what I mean by that, why it's important, and given you a glimpse of how it might work. But first, I need to set the scene. Let's look at a day in the life of a 21st century knowledge worker. I arrive at my office and look at my inbox full of emails. Let's summarize it. Okay. I'm struggling to figure out how to respond here.
[00:19:14] So let's get AI to write a response. Next, I need to write a report. But I'm struck by the blank page problem. I know, I'll drop in some resources and get an AI draft. Looks good to me. By the way, the writer's block used to be staring at a blank page. Now it's staring at a page that AI filled out for me and wondering if I agree with it. I've become a professional validator of a robot's opinions.
[00:19:43] I've got some data to analyze. Maybe AI can analyze this data for me? Probably correct. Okay. I've got to make a deck as well. You know the drill. All right. Oh, I was supposed to prototype something as well. Ah, okay. Let me vibe code something. All right. All this looks good. Let's go. This isn't a vision of the future.
[00:20:08] This is a completely plausible, if slightly exaggerated picture of the world of knowledge work today. Welcome to the age of outsourced reason, where the knowledge worker no longer engages with the materials of their craft. We've become intellectual tourists. In our own work, we visit ideas. We don't inhabit them. Our relationship to our work is entirely intermediated by AI.
[00:20:37] Some might say alienated. We've heard that story before. What I want to focus on today is that using AI in this way can have profound implications on human thought. Consider creativity. On an individual level, we might think that AI is a creativity boost, giving us rapid access to new ideas.
[00:20:58] But numerous studies have shown that on a collective level, knowledge workers using AI assistants produce a smaller range of ideas than a group working manually. We've created a hive mind. Except the hive is really boring and keeps suggesting the same five ideas. Consider critical thinking. We surveyed knowledge workers about their use of AI.
[00:21:24] They reported that they put less effort into critical thinking when working with AI than when working manually. And this effect was greater when they had greater confidence in AI and less confidence in themselves. Consider memory. When people rely on AI to write for them, they remember less of what they wrote. And when they read AI-generated summaries, it's hardly surprising that they remember less than if they'd read the document.
[00:21:54] And finally, consider metacognition, which is the ability to think about your own thinking process. Working with AI requires significant metacognitive reasoning about your task goals, decomposing the task, the applicability of Gen AI, your ability to evaluate the output. These are things which are built into the process of working directly with a material and which become problematic when that material engagement becomes intermediated. Basically, we've become middle managers for our own thoughts.
[00:22:25] So what's the score? We have fewer ideas. We think about them less critically. We remember them less well. And we have a harder time doing it. Taken together, we can see that AI-assisted workflows can have profound effects on human thinking. And this extends even to seemingly trivial, mundane tasks, because these everyday opportunities for exercising our creativity, our critical thinking, and our memory are essential for protecting our cognitive musculature
[00:22:54] and allow us to rise to the occasion when an exceptionally complex task comes our way. Studies show that when we don't use our brains, they get worse at brain things. Nobel Prize Committee, please hold your applause. Is this the cost of progress? We've solved the problem of having to think. Unfortunately, thinking wasn't actually a problem.
[00:23:22] It's like we invented a cure for exercise and then wondered why we're out of breath all the time. It doesn't have to be this way. Beyond AI as an assistant, I believe that AI should be a tool for thought. AI should challenge, not obey. And I believe that right at this moment, we are at a critical juncture, where the world of work is poised to be transformed by generative AI.
[00:23:48] And we must act now to shape and drive that transformation towards humanistic values. Of these two diverging roads, we must take the one less travelled. Beyond getting the job done, a tool for thought helps us better understand the job. Beyond getting it done faster, it helps us get it done better. Beyond getting us to the right answers, a tool for thought helps us ask the right questions.
[00:24:16] Beyond automating known processes, it helps us explore the unknown. What does this look like? What I'm about to show you is a prototype developed by my colleagues and me at the Tools for Thought team at Microsoft Research in Cambridge. Now, please bear in mind that this is a live research prototype. It's not a product. And it's just one of a series of explorations that our team is conducting to study how different modes of working with AI can enhance human thought.
[00:24:44] So let's look at a fictitious example. Clara and her colleagues run a company that sells bottled beverages. They've just had a meeting to discuss a new industry report that seems to have some pretty important findings about consumer preferences for sustainable packaging. Clara's colleagues have asked her to write a proposal arguing for how the company ought to respond. So she really needs to get to grips with this report,
[00:25:14] understand its findings and its data and how it fits into her business context. She starts by loading some documents into her workspace. There's the meeting transcript to remind her what was discussed. There's a recent internal report from her own business. And of course, there's the industry report which she opens. She sees an overview of the document along with section-by-section summaries. Except these aren't really just summaries.
[00:25:42] We think of them more as lenses. They're customizable micro-representations of the text that can emphasize what is most relevant to the task at hand. So in this case, Clara selects the consumer's lens. She can select a section for deeper reading. As she reads, she makes notes about her thoughts and highlights excerpts from the document. As she reads, she also sees AI-generated commentary and critiques.
[00:26:12] We call these provocations. Note how this process is a hybrid of completely manual reading and completely relying on AI to read for you. Clara still reads, but intentionally and strategically. Now as Clara is working, she's building up an outline of her argument manually. This outline is lightly structured and allows her to sketch out the flow of her argument at a high level while still retaining deep connections and being grounded in the source documents.
[00:26:41] As a result of which, we can already generate a draft of the proposal. And Clara can do things here like add a heading to the outline to generate a paragraph. But what I want to draw your attention to here is that while this text is AI-generated, Clara has a completely different relationship to this text than if she just dropped in some documents and said, write me a report. Because this text is deeply rooted in a cognitively effortful but interactionally effortless thought process.
[00:27:11] It reflects Clara's decisions, Clara's judgments, Clara's unique personal professional expertise. She sees another provocation, this time in the outline. In this case, she decides that while the provocation is useful, she does not need to address it. Unlike typical AI suggestions, provocations are not meant to be applicable all the time. They're instead meant to stimulate your thinking about your work. Because if you understand your work well enough, deeply enough,
[00:27:41] to make the confident decision not to accept a piece of feedback, then the feedback process is still working as intended. But we're not done yet. Clara has entirely new ways of interacting with this text because of generative AI. A really simple example is that she can just resize a paragraph to change its length. She can also rapidly test different versions of this text. And at select strategic points, indeed, she writes.
[00:28:10] As she writes, she sees provocations that, rather than auto-completing her ideas, they raise alternatives, they identify fallacies, they offer counter-arguments to help her strengthen and develop her own argument. There's something you won't find anywhere in this interface, and that's a chat box. Clara's not having to chat with anything to do her work, yet she is silently and appropriately assisted by her computer as a computer and not as an ersatz human.
[00:28:40] Throughout this process, Clara has been assisted and, yes, probably worked faster because of AI. But she's also maintained direct material engagement at strategic points. She read the relevant portions of the document herself. She constructed her decisions and her argument herself. And ultimately, it can be said, she has written this document herself. Moreover, she worked better because of AI. AI provocations at every stage of the process kept her metacognitively engaged,
[00:29:09] always looking for critiques, alternatives, and lateral moves. We have been studying the effects of tools like this, and the results are promising. You can demonstrably reintroduce critical thinking into AI-assisted workflows. You can reverse the loss of creativity and enhance it instead. You can build powerful tools for memory that enable knowledge workers to read and write at speed with greater intentionality
[00:29:38] and remember it, too. It turns out, with the right principles of design, you can build tools that are the best of both worlds, applying the awesome speed and flexibility of this technology to protect and enhance human thought. These are simple, general principles like ensuring that the tool preserves material engagement, offers productive resistance, and scaffolds metacognition. And while we've been primarily studying
[00:30:08] professional knowledge knowledge workers, we believe that these principles can extend to all aspects of AI use, including when we use it in our daily lives, our hobbies, and even in education. I repeat, efficiency is not the aim of tools for thought. Better thinking is. But sometimes you can't have both. I used to think there was no such thing as a free lunch in human thinking. This is so much better than a free lunch. This is a lunch that pays you to eat it.
[00:30:40] I want to close with some thoughts on the values that we have in developing AI software. What if AI gets to the point where it can do a better job of thinking than humans? Why should we care so much about protecting and augmenting human thought? There's two reasons. First, there may always be ways of thinking that remain unique human strengths of which we may not even be aware. Second, perhaps more importantly, we take the position that the ability to think well
[00:31:09] is essential for human agency and empowerment and flourishing. This echoes an ancient question. People once asked if writing, if books, if the internet can remember for us, does it matter that we cannot? People once asked if maps can navigate for us, does it matter that we cannot? Now we ask if machines can think for us, does it matter that we cannot? If machines can speak for us,
[00:31:38] grieve for us, pray for us, love for us, does it matter that we cannot? To me, the answer is pretty obvious. When I began studying human-AI interaction 13 years ago, it was inconceivable to me that we would be asking these questions in my lifetime. But we are, and we must. I leave you with this thought. What would you rather have? A tool that thinks for you or a tool that makes you think?
[00:32:10] That was Advait Sarkar at TEDAI Vienna 2025. Outsourcing our vernacular, the very texture of our language to AI might help us defeat the dreaded blank page, but at what cost to our cognition? I'll admit, when I'm pressed for time, I'm pulling up a cold pilot to help me fire off messages, analyze reports, or structure and outline. It's all in the name of efficiency.
[00:32:39] Back when I was working in big tech, efficiency was the north star. But in our rush to automate the busy work, we may have overcorrected and automated the process of thinking itself. We've bypassed the productive struggle of ideation. And I get it. Ideation can be a difficult and frustrating process. But we need the friction from grappling with a thought, hunting for the right pros, and nurturing a seedling of an idea. Without this process, we limit ourselves to a smaller range of ideas
[00:33:09] we remember less and think less critically. This doesn't mean we have to completely abandon AI. Instead, we must reframe how we want to use it. Instead of helping us work faster, we can use this tool to think better and enhance our creativity. All right, that's our show. Thanks for listening. TED Tech is a podcast from TED. This episode was produced by Rahima Nasa. Our editor is Alejandra Salazar,
[00:33:39] and the show is fact-checked by Julia Dickerson. Special thanks to Constanza, Gallardo, Daniela, Belarreso, Maria Ladias, Tanzika Sangmanivan, and Roxanne Heilash. If you're enjoying the show, make sure to subscribe and leave us a review so other people can find us too. I'm Sherrell Dorsey. Let's keep digging into the future. Join me next week for more.