Episode Transcript
[00:00:05] Welcome to Prompting Curiosity, a podcast for the AI curious. No coding background required. I'm your host, Dr. Shantae Cofield, also known as the Maestro, and I created this show to explore what these AI tools actually are. Really, though, are the files in the computer, how to use them, and what they might mean for how we think, work, create, and move through life. Whether you're skeptical, intrigued, or already experimenting, you're in the right place. All that I ask is that you stay curious. All right, let's get into it.
[00:00:38] Hello, hello, hello, my curious people. Welcome to episode 47 of Prompting Curiosity. I'm your grateful host, the Maestro, and today we are talking about how people are actually using AI in 2026. Right? It's super easy to think that everyone else is doing what you're doing. Spoiler. If what you are doing is coding, then basically no one else is doing that. Um, um. But I figured I'd take a look at some numbers and together we'd take a look at some numbers and see what they say about how folks are actually using AI. So real quick, before we get into that, and speaking of coding, I, uh, just want to. Want to. To thank anyone, thank everyone who attended last week's intro Vibe Coding workshop. Um, it was awesome.
[00:01:22] Uh, and if you were there or you watched the record, if you watch the recording. Bless you. Um, but if you're there, just big, big thank you. Um, for those wondering, we had seven folks, uh, sign up, so I limited it to ten, and that was the stretch. I knew that just based on the initial interest and just the numbers of what. What we're going to talk about going forward, like, support it. It's not a. It's not a widespread use case.
[00:01:43] Uh, so I knew that 10 would be a stretch. We got seven folks, and, uh, I think five showed up lot four or five showed up live. So if you were there, amazing. And thank you. I am not sure if I'll run it again. I'm just really about the numbers. Right? It is a niche, niche usage. Uh, uh, and the data shows that. But if you're really wanting to learn and you're like, I couldn't make it, whatever, I don't like the replays, whatever.
[00:02:10] And, uh, you want to learn how to use AI to vibe code, reach out and I will see about running another if there's enough interest. Um, but in reality, it'll probably be that I sell the recording or maybe rerecord me. Be like, rerecord turns to a course. I don't know. My hesitation with recording things is that things change and then I don't have to go change the course. So it's easy to run it live. But if you're really interested, reach out, as I do want to help everyone that wants help with it. And, um, we'll go from there. Right, so let's get into the episode how folks are actually using AI in 2026. So the data. Where did I get this from? Uh, Claude, my French assistant, and I, we used five sources, uh, for this episode.
[00:02:48] Uh, first I'm gonna go through them and just kind of highlight like, what they are and, you know, any possible limitations of this. It's just, it's good practice. Right. So first was a. It's a study, AI in the Wild study, published in Harvard Business Review. It was just published recently on June 1st. Actually.
[00:03:03] This, um, is the third annual installment of a longitudinal research that analyzed 12,637 AI use cases collected between March 2025 and February 2026, um, across Reddit, Quora, LinkedIn, TikTok, YouTube and other articles. Um, this methodology, Fordsworth, is called social listening, meaning the researchers are capturing what people voluntarily share publicly about how they use AI. It's not like a controlled survey, but worth noting. Obviously, people who actually post about their AI usage, they are. And they post about it online. They're skewed towards this, you know, a heavier, more engaged user. The average person who's using ChatGPT to like, look up a recipe likely isn't posting about it. So it's worth noting. Second Source was an OpenAI and National Bureau of Economic Research paper. This analyzed 1.5 million actual ChatGPT conversations. So it's a good sample size. Um, but a big caveat Here it is OpenAI setting their own project, which obviously creates incentive to frame things positively. Um, they also only looked at the consumer plans, not the business accounts. So. Interesting. Third, Stanford 2026 Human artificial intelligence Annual Report. This pulls from a variety of sources. Um, and this is a little opposite of the other things. It focuses on individual instead of, um, it focuses more on things like technical capabilities, adoption rates, workforce, uh, impact, and public perception. So a little bit of the opposite side side of things. Right? Um, fourth source was the 2026 Organization for Economic Cooperation and Development.
[00:04:41] Uh, if you read this, then it would be. The acronym is oecd. Um, and so is their data. This leans towards workforce and economic metrics rather than what people are actually doing day to day. So I wanted like a well rounded, you know, information sampling of things. And then last but not least, this is Pew Research center, um, their data, they are a nonpartisan, non advocacy polling and research organization. They are based out of Washington D.C. they have been tracking American attitudes towards AI for like five years.
[00:05:07] Um, and they do this through ongoing surveys of US adults and teens. Caveat here, the Pew data, like I said, it is American attitude, so it's US specific. Um, so it doesn't generalize globally the way some of the other sources will, but still worth having in here. So yes, I asked Claude to help find these sources and Michael was to just balance things out and get as well rounded of a picture as possible. So let's get into what the data is saying. So who is using it? AI. We know adoption is growing. More than a third of individuals across the OECD countries use generative AI tools in as of 2025. So worth noting here, OECD, that's made up of 38 wealthy developed nations. So we're thinking like Western Europe, the United States, Canada, Australia, Japan, places like that. Um, so it's, it's only, it's limited to those people, but adoption growing within that demographic.
[00:06:01] Um, 31 of the people survey said that they interacted with it at least several times a day, which is up from 2020, from, which is up from 22%.
[00:06:10] And um, that was in early 24, 2024. So up front, we're currently at 31 survey saying they're using it, uh, several times a day versus 22% last year, two years ago.
[00:06:21] Um, the most significant demographic gap is actually age. So around half of adults under 50 that were surveyed say that they interact with AI about once a day or more.
[00:06:34] Roughly two thirds of US teens ages 13 to 17 say that they use AI chatbots. And then three quarters of students 16 and over, um, um, this is across those OECD countries, they report using generative AI tools. Right? So we know this actually as well because when they look at usage, it plummets in June, or it historically has plummeted in June, and that's because teenagers are out of school and they're not using it anymore. So this data tracks with, if you're in the AI world and like following any other stats, like it tracks, you're like, yeah, the majority of people using it, or we should say more people of the younger teens demographic is using it than, uh, people over the age of 50.
[00:07:18] All right, um, so worth noting just in terms of like this, you know, the continuation of AI and, and the perpetuation of it.
[00:07:28] Um, so yes, next thing as it relates to demographics, um, according to oecd, and contrary to what the AI shills would have you believe, the gender grab is apparently relatively small. So the OECD puts it at 4.2 percentage points, um, with men using it slightly more than women. We're seeing some of these celebrities come out and being like, women are getting. It's female celebrities coming out and being like, women are getting left behind. And it's like this. Some AI company is clearly paying them to sa these things because they never talked about before, right? Reese Witherspoon's one of them. And you're like, I love you, Reese. Why the are you doing this? All right, money makes people do crazy things.
[00:08:02] Um, but the data is showing us that it's not actually this massive of a, um, as massive of a gap as people would think. Um, but the data in general is thin and the methodologies they used to track it outside of the OECD, um, data, right? OpenAI had some data published, but the way that they were determining the gender and the sex of the, of the user was based on the username. And so like, that's like, not super accurate. But just to say that from the data we do have, it isn't like this massive divide.
[00:08:36] Next, the. Which was what the title of this episode is, what are they actually doing? Right? What are people actually doing? So the OpenAI paper that I spoke about or that I introduced, um, they broke usage into three behavior types. About 49% of interactions are asking, right, using AI as an advisor, getting explanation, exploring ideas. Um, and that is the fastest growing and Highest rated category. 40% is doing so that's drafting, planning and task creation. And then 11% is expressing, which they categorized as personal reflection and play.
[00:09:11] Um, writing is the most common work. Task coding and self expression remain niche for everyday users. And I highlight this because I did a whole workshop on, you know, vibe coding. And I, I know like, it's not. It's very niche and we put ourselves in these silos and like, I'm on threads and you think they're like, everyone's doing it and it's like, no, no one is doing it, right? Most people are just hanging out with, with AI.
[00:09:35] So the Harvard Business Review, the, uh, the study that was published in Harvard Harvard. Wow. The study that was published in Harvard Business Review, um, that AI in the wild study, they ranked the top use cases from those 12,637.
[00:09:49] So the top 10 use cases in 2026 are as follows. Number one. Can you guess, folks? Number one therapy and companionship. This held the top spot for the second year in a row and grew from 5% to 11% of all use cases. It grew 6% or whatever that is. 6. I don't know the actual math on that, but it went from 5% to 11% of all use cases in a single year. So that's, that's what it doubled. So it's not 6%. You guys understand what I'm saying there? You folks understand what I'm saying?
[00:10:25] It went up by 6 percentage points.
[00:10:29] Number two, use case troubleshooting. Number three. I don't know exactly what's entailed in this, but fun and nonsense. So I guess when you have it make a cow on a skateboard or write a poem in the, you know, voice of someone else. Number, uh, four. This was interesting. And this is a new entry this year. And like, this is a use because I like literally would have never thought of, um, which is fan fiction and storytelling. So it just, I guess you have it write a story.
[00:10:56] Uh, and if you want fan fiction, you know, you can have it ship, uh, you know, two of your favorite characters and now they're in a relationship and you can have it right Stuff.
[00:11:04] This is smart. I would have never tried to do that.
[00:11:07] Um, smart is a way to think of personal entertainment. Number five, technical software use. Number six, autonomous agentic operations. Right. So we talked about that a bunch. An agentic, uh, AI. So this is when AI is actually doing things rather than just advising. This is a new entry this year and it makes sense. I, I bet the number one spot here is of, you know, companies that are, are seeing this. It's going to be going to be clawed, but we're going to be anthropic. But. And it didn't give that specific data.
[00:11:35] Number seven.
[00:11:36] I bet you can guess this one. I feel like we're on Family Feud. Is that what the is a family Feud? Is that the. Yeah, no, I. Give me the top ten. Um, either way. Number seven is relationship advice.
[00:11:48] Number eight, work buddy. What that means? Don't know. Uh, but that's what, that's what they categorize it as. Number nine. Interesting. And a new entry, astrology and tarot readings. And then number 10, general advice. So three things worth noting in that list.
[00:12:08] Three of the top 10 are brand new this year.
[00:12:12] Right? This, this the way people are using it. It's expanding, it's changing. And literally three of the top ten brand new self improvement and reflective uses actually fell out of the top 10. People are not trying to better themselves with, with AI. Right.
[00:12:30] And last, the emotional support category is the single largest and growing the fastest.
[00:12:38] And, um, we know this, I think intuitively.
[00:12:42] I'm, um, not surprised by this at all. So there's your top 10, um, uses. I still have a few more things I do want to talk about. I'm just kind of diving a little bit deeper into, into things. Um, but as for how people in general are using it, there you go. I don't know where you fall in that list, but there's our overview. So something that I wanted to dive into, but there honestly wasn't much good data yet. Um, is it is or was is AI being used as a search tool replacement? Uh, and most of what's being published on AI versus Google search behavior, it's coming from SEO companies and marketing agencies. So these are sources with a pretty obvious, obvious financial interest and how that story gets told. So, like, I don't take their data as that helpful.
[00:13:35] Um, the one study that we have from our, from our original list of resources is Pew, and they tracked the actual browsing behavior of 900 U.S. adults in March of 2025. So small sample size and U.S. only. And it's from last year. But what they found was 58% of the users encountered, uh, an AI generated summary during a Google search. And when that summary appeared, they clicked a traditional link only 8% of the time versus 15% of the time when no summary appeared. So I know it's not a massive difference. Um, but when AI, for those people, when AI gave them the answer up front, they were roughly half as likely to click through to a traditional link. So, you know, you get that little AI summary and you're like, all right, cool, there's my answer. But for those people, um, what I took away from reading, you know, I did read through the less reputable sources and what I took away from it, and it does track with my own personal behavior, is that people are going to AI first for some searches. Some things from there they may then also go to Google to verify. So ultimately, to me it feels like a hybridization of things more than a replacement of anything. All right, next, uh, category here. Personal work versus or personal usage versus work split. So we know like the majority of much of the way that people are using this is personal. But let's talk about what's happening at work as well. So first, just the, just to name and give some data around the split at work. About 1 in 5 US workers say that at least some of their work involves AI, which is up from 16%, uh, in 2024. Um, but a majority, 65% still say that they don't use it much or at all. So again, it's not everywhere, it's not ubiquitous, it's not like just taking over everywhere. Um, but it is increasing.
[00:15:28] According to OpenAI's own research, approximately 70% of ChatGPT consumer usage is personal and 30 is work related. Right. Which both are continuing to grow. But like, it tracks with what we saw for the top ten things.
[00:15:43] Um, and I think it's worth flagging is like, I don't know how they like get this data. All right, because they're basing enough of conversations. And I'm like, so this is of the conversations that they're allowed to look at. I don't really know. I, I doubt it. I feel like they'd just be doing illegal. But like it is something to think about. Right. Either way, we're looking at the personal versus the work split.
[00:16:03] We know that the main use case is not productivity, uh, even though that's the story that dominates headlines. Right. People are mostly using AI for their personal lives. I.
[00:16:14] So, you know, you listening to this perhaps, you know, continue to be in the minority. Um, because I use it really heavily for work, I'm not using it for personal stuff, uh, that much.
[00:16:24] I will chat with it from time to time, I'm not gonna lie. But the majority of the way that I'm using it is for efficiency and productivity and to make my life, my work, the work side of my life, easier and faster and allow me to do the stuff that I like, want to be doing.
[00:16:38] Um, so let's talk about AI usage at work. According to the Stanford index, 58% of employees globally reported using AI at work on a semi, regular or regular basis in 2025.
[00:16:51] That number varies vastly and dramatically by region. Um, whereas we see in India, China, Nigeria, UAE, Egypt and Saudi Arabia, it exceeds 80%.
[00:17:02] So again, even within these two studies. Right. So earlier I said that, that uh, about one in five US workers say that some of their work involves AI, right? So that's 20%.
[00:17:15] Whereas they look at the standard, the standard, um, AI index and this is looking at globally then 58% of, of workers or employees globally are using AI work. So again it is regional and things like that. Um, but again I'm just gonna, I'm just trying to give you numbers and see like, yes, people are using it at work. It doesn't seem to be the mass adoption and uptake that these companies have would have you believe. But yes, people are using it at work.
[00:17:43] Um, the Harvard Business Review data shows that most workplace use is generating efficiency gains. Right. It's summarizing, drafting, saving time. Um, it's not fundamentally transforming how work gets done. And that is worth noting. It's not replacing. Right. It is assisting, augmenting.
[00:17:59] Um, a fun one here and a new term to me, shadow usage. Uh, and that's when you use it, you don't tell people about it. Um, that is apparently widespread and that is people using AI without telling their employers. Um, and, you know, quietly getting the work done faster, keeping the gains to themselves.
[00:18:17] This is a thing, um, which makes sense. There's probably more friction at large organizations, right? The IT restrictions, governance concerns, reputational risk.
[00:18:26] Um, so people are tiptoeing around these things. So that is a new thing, a new term. Um, but in terms of replacement, we know it's a thing. And then I think that we're going to see a shift. But, um, among entry level workers. Employment. Employment. Wow. Among entry level workers, employment for software developers age 22 to 25 has dropped nearly 20% since 2024. Um, even as older colleagues headcount does grow, the pattern repeats in other jobs with high AI exposure, like customer service. And I spoke about this m. In an episode I did, you know, AI is not coming for your job, it's coming for some people's jobs. Absolutely. And I do think that we will see course correction. I think we'll see a reversal of things. Um, but yes, initially it is. And when in that episode, what I'm talking about is for those of us in the service industry, those of us that are really in the business of trust, it is not coming to replace us. But for folks that do have AI exposure, yes, it is coming for them. And I did use an example about customer service because that was really one of the first domains to get hit and really significantly impacted by AI. But that is one of the first company, you know, domains to see that reversal. And seeing people being rehired after these, you know, the robots are not doing what. What they're supposed to do. So that's a good segue into the next, uh, and the last section here, which is the concerns. Right. So according to Pew, half of us adults surveyed say the increased use of AI in daily life makes them feel more concerned than excited. We know you can feel this. You can just read it online and you're like, people are not stoked about this. On the whole, right? Only 10% of the people, ah, they tracked said that they're more excited than concerned. Um, about half of Americans surveyed say that AI will worsen people's ability to think creatively, um, and worsen their ability to form meaningful relationships. Absolutely. And this is one of the first things that I did as one of the, one of the first episodes. Right. Talking about it. You're making, you know, our intelligence, our creativity, the way that we socialize. I do believe that if you just outsource everything to AI, yeah, it will cause a problem, but it doesn't have to cause a problem. Right. You can use it to augment things instead of replace things. But, yeah, you know, just defaulting to it. Absolutely. Um, and so speaking of that Harvard Business Review study introduced a term called think slop. And so the author of that paper is Mark Zhao Sanders. Um, and he. This is his term and the, the categories. Categories that he introduces. Um, it's his term. Think slop is his term for cognitive laziness.
[00:21:03] Uh, that's enabled by over reliance on AI. And the, the four behaviors that he kind of, uh, lists out associated with this, um, they're his, they're like his hits, his framing. But the data does support what he's naming. Right. So the four behaviors of think slop, one, losing track of your intention before prompting. Two, outsourcing your thinking entirely. Three, stopping writing altogether. Right. And writing is a form of thinking. Uh, and then four, developing false intellectual confidence because AI sycophancy makes everything feel better than it is.
[00:21:39] You can just read that and like, yes, I, I see this, I understand this, I acknowledge this. I agree with this.
[00:21:45] Absolutely. Absolutely. Does it have to be the case? Can humans think independently and make their own choices? Yes. But if you just outsource everything to AI, does think slop become the outcome, the logical outcome? Yes, of course. So the takeaway here, uh, is that people, at least to me, is I think the biggest takeaway that people are largely using AI for personal reasons. Right. With therapy and companionship coming in at number one to me. Right. This is.
[00:22:18] This causes me to raise an eyebrow.
[00:22:21] Right. More than think slop. Right. This causes me to raise an eyebrow because while the public is clearly torn about AI, Right. The folks who are using it are using it in a way that I don't think they will ever be willing to give up.
[00:22:37] Right. The companies and the tech bros who stand to make all the money, they're super bullish on it, of course, and they are shoving it down our throats and they're hitting the productivity angle. Right. They're really hyping that productivity angle. And they're threatening with this, you know, adapt or get left behind. It's inevitable.
[00:22:56] But what we're seeing is a rehiring of humans as AI is not living up to the promises and a rehiring of humans as CEOs start to get the actual AI bills as VC funding runs out.
[00:23:10] So when it comes to the future of AI, what I think deserves more attention and more discussion is this personal usage and the very real likelihood that the quote unquote fight won't just be between people and AI or people and corporations, but also, and perhaps predominantly between people and the other people who are defending the AI.
[00:23:37] We saw it when OpenAI tried to retire GPT4O. We saw that pushback big time. Right. Uh, people weren't upset about losing a productivity tool.
[00:23:48] They were upset and they were willing to fight because they thought they were losing and felt that they were losing something that felt personal.
[00:23:56] Big, big difference. And I think something that deserves a lot more attention. So we'll see.
[00:24:02] We will see. All right, looking at the time, let's wrap this thing up. How I used AI this week. If you're new here, welcome. Each week I share a quick example of how I used AI that week. So this week I used it to help me vote. Don't worry, folks, don't worry. So I live in California and we had our primaries, uh, and there were literally 11 billion candidates, a zillion different positions. All right, um, FYI, I do not vote for mayor of LA because I do not live in LA. I live in LA county, but I do not live in LA City. Uh, but I would throw myself in traffic before voting that whack ass piece. Voting for that whack ass piece of the dispenser Pratt immediately. No, um, I do, however, vote for governor and for me it was Theo Becerra all day, all the way. So claps to him, um, the rest of the positions. That is what I had Claude help me with. So don't worry, I did not ask Claude who to vote for. I asked Claude to pull sources and then aggregate the information and give me summary of the candidates and their positions and their, their actions and things like that. Um, my way of checking this was to then ask it to also provide me with the sources that it used as well. And I, I did also submit some of my own sources. Um, the process still took long as, like, it is not ever a short thing to vote here.
[00:25:14] Um, but I was pleased with it and pleased with what what Claude produced um, and was able to help me with. So that was my use case. So that my friends is all for today. Hopefully you found this episode helpful.
[00:25:27] There's a longer one so thank you. Um, if you did consider sharing it with someone who you know is curious about AI right. The more the merrier over here. We want an intelligent, informed and curious crowd.
[00:25:39] Don't forget I have a companion newsletter and blog, the Computer. Wow. The computers, the Curious Companion.
[00:25:45] Uh, and both drop every Thursday and are the podcast episode in text format. So if you prefer to read or you want a written record, join the newsletter or head to the blog, you can head to prompting curiosity.com forward slash newsletter or prompting curiosity.com forward slash blog or just very simply check out the link in the show notes as always, endlessly, endlessly one more time. Endlessly appreciative for every single until we chat again next Thursday.
[00:26:14] Stay curious.
