We touched down in Orlando on the morning of Saturday the 27th of June. Daniel had taken a redeye from San Francisco, and I had woken up at 4:30am to pack and take a flight from New York where I’ve been spending the summer as a research assistant. We met up with each other at Gate 43 of the Orlando International Airport, where I got off my flight at around 10am local time. We dapped each other up and shared our collective awe that this trip, which had been a mere wisp of an idea just a few weeks prior, was now very real. And also very hot. Orlando was burning when we touched down, and the humidity definitely didn’t help. When I stepped out of the airport, it felt like I had been hit with the breath of a particularly ferocious dragon. We then ordered our Uber, and proceeded to the hotel where we would be spending the next four nights.

As a native Californian, I am ashamed to say I had a bit of a preconceived idea of what I thought Florida would be like before I arrived. I expected the heat, obviously, but I mostly expected other negative things. In my mind, I pictured exasperated people, miles of dry land, and landscapes that were at best underwhelming. What I experienced, however, was perhaps the exact opposite. From the moment I touched down to the moment I took off, I was met with with the kindest people, the most beautiful palm trees, and the most lush, green scenery that I can recall in recent memory. The heat and tropical lightning storms were awful, yes, but the Floridian hospitality I received most definitely made up for them.

Now, the conference itself. When Daniel and I were checking the prices to rent booth equipment a few weeks prior, we nearly fell backwards off our seats. More than $1000 for a chair?!?! Did these chairs contain magical powers that I was unaware of?? Regardless, we realized that renting directly from the conference center was most certainly not the move. We ended up heading to a local Orlando Target and buying foldable tables, chairs, stools, rugs, and bowling pins that we planned to return after the conference. Even when we were transparent to the Target employees about our plan, they kindly allowed it and even smiled at our shrewdness (this is the Floridian hospitality I am talking about!). We ended up lugging (to our Uber driver’s dismay) all of the newly purchased equipment to the conference center, where we were met with another misfortune: because we weren’t 18 yet, we would not be allowed on the Expo Hall during setup, as it was a liability to the conference in case we got injured. Fortunately, due to the combination of kind employees and smooth talking from Daniel, we were allowed to leave our heavy tables and chairs by the registration center and set it up as the conference hall opened to the public on Monday June 29th. With our tails slightly between our legs, we returned to our hotel room, but at least without the heavy equipment weighing us down.

Sunday the 28th was a blur. Daniel and I had a great dinner the night before, and planned to wake up at 9am to plan our conference out and get ready for the 4:30pm Welcome Reception that ISTE was hosting. Unfortunately, being the undisciplined teenagers we are, we slept through all five of our alarms and, to our shock, found the time to be nearly 1pm as we regained consciousness. After a Hawaiian bowl we hastily ordered on Doordash, we started our planned outreach. About one hundred educators and technologists who had posted about ISTE received annoying messages from Daniel and I at this time, and while many probably didn’t even see them, we felt proud of our LinkedIn handiwork. We decided to reward ourselves with a nice walk around the resort, which we completed before we headed to the 4:30 reception. I realized to my horror at the entrance that I had left my badge back at the hotel room, but a kind ISTE employee who had met us the day before just laughed and printed me out a new one.

The welcome reception went, in a polished way, not as well as expected. In a perfectly blunt way, it was pretty shitty. It was both Daniel’s and my first time at an open-networking event like that, and everyone we saw already seemed to be deeply immersed in a conversation. While we did talk to a few people, including the founder of a Mexico-based EdTech company, we found our performance to be a disappointment. Heading back to the hotel with our tails between our legs for the second time in two days, we decided to preserve morale by calling the event a “learning experience” and putting it behind us.

The first real day of the conference was Monday, June 29th. We got to the conference right around 9am when the Expo Hall opened, and scrambled to assemble and organize our tables, chairs, and electronics while the rest of the teachers and educators were let in. The first 45 minutes of our presentation time turned out to be just setup, as we had to unpack, prop up, and carefully place all of our equipment, all while making sure to meticulously collect the tags of each item so that we could return them back to Target after the conference ended. When we were finally set up, we began to get in business, and our first few hours ended up going really well. We also met our booth neighbors, Filewave, who were some of the kindest people I have ever met. Not only did they allow us to share their power that they had paid for, but they also were just extremely good people (another point for Floridian hospitality). It is rare that you meet people who truly go out of their way to help total strangers (and to an extent, competitors as booth neighbors), but that is exactly what they did for us. Thank you deeply to Lena, Lisa, and the other kind Tennessean man whose name is evading my brain at this moment. For the history of Deskpad’s existence, we will be indebted to you for the kindness you showed us at this conference.

The rest of day one went by just as quickly. We had conversation after conversation with teachers, educational technologists, and open-minded attendees about what exactly we do at Deskpad, and how they could support us. The throughline was that many were just impressed that we were doing this while in high school, and we received an overwhelming amount of support and praise, which we were extremely grateful for after two underwhelming days so far. Being the scrappy and forgetful team that we are, we realized we forgot to buy the software that could scan people’s badges and put their information into a built-in CRM, so we decided to go the old-fashioned way and collect people’s names and email addresses in a blue flower-covered notebook we had purchased as an afterthought at Target. We ended the day extremely exhausted from repeating our pitch over and over, but felt extremely gratified to have been given this opportunity. Here’s a quick photo of what we looked like:

Around 6:30pm, we returned to our hotel room, exhausted and ready to sleep. However, we looked at the weather forecast, and realized that we would not have the opportunity to use the hotel’s grandiose pool if we didn’t go tonight, as the next couple nights would be filled with lightning storms. So, despite being exhausted, we headed down to the hotel pool for what ended up being an extremely fun time. We returned to our room sopping wet and, after a quick shower from both of us, decided to order in Chipotle and omit another late night out. Unfortunately, despite our seemingly logical line of reasoning that if we stayed home and ordered food we’d go to bed earlier, we ended up starting the Jason Bateman-starring Ozark, which we watched late into the night.

We woke up at 7am for day two of the conference extremely sleep deprived, but we pushed through. Day two’s first half ended up being pretty slow, as the heightened excitement and novelty of the Expo Hall from day one had worn off a bit. However, about halfway through the day, we had a brilliant idea. Rather than passively sitting at our chairs and waiting for interested passersby to come to us, Daniel would wait out in the front and ask every person who walked by the same thing: “Hi, we’re a couple of high school students here showcasing our project. Could we steal two minutes of your time to tell us what we do?”

It worked like a charm. Though I had told Daniel this idea would have a 90% conversion rate, I mostly said this to get his morale up and to get him to actually put himself out there. However, even I was surprised to find that my then-optimistic prediction was actually quite accurate. Nearly every onlooker’s ears perked up when we mentioned our age and mission, and they were even more impressed when we showed them what we do. The second half of day two had hands flying as our booth visitors scribbled down their name and contact information in our blue-flowered notebook. We also received a couple of compliments on our pen—a dull baby blue Japan-made one I found in my brother’s bedroom. Day two was a massive success, and we came home again exhausted but again triumphant.

I won’t write too much about day two’s evening because to be honest, we really didn’t do anything. Exhausted from our late night the night prior, we didn’t have the energy to go out, but also we didn’t have the discipline to go to sleep on time. What ended up happening was another night of Doordash—this time Cuban food for me and Panda Express for Daniel—and a 2:30am bedtime fueled by how good Ozark was getting. When we finally shut our eyes, we realized we were screwed for the next morning.

Day three—Wednesday July 1st—felt like it arrived far too soon for our sleep-deprived brains and eyes. However, the one thing I had been harping to Daniel about all trip was the continental breakfast buffet at our hotel. I love food. I am a foodie, I admit it. And hotel breakfasts are like the Holy Grail for us foodies. You have mountains of eggs, bacon, pancakes, and every pastry you can imagine lining the kitchen, and you can go back and get more as many times as you’d like. Despite being extremely tired as I shut off my 7am alarm, I had to nearly shake Daniel awake so that he would join me at my breakfast paradise. Even on a cumulative total of less than 10 hours of sleep in the past two nights, the breakfast was awesome. I ended up getting three plates, and though my eyes ended up being a bit bigger than my stomach, I had a fantastic experience. It felt like a very fitting way to leave our hotel, and we checked out officially at 8:45 and headed to the convention center for our last few hours of presenting.

On our third and final day, we continued the strategy we had discovered the day prior of Daniel standing outside the booth and attracting onlookers. As we were walking in, Daniel optimistically predicted that we would fill two pages of our notebook with names and emails due to our new strategy. We ended up filling more than three. Day three was a fantastic experience, and in the last hour of the conference, we decided to close up a little early and actually explore the other booths as visitors. We went to the big dogs, obviously, like Google, Microsoft, and Adobe, but our most enjoyable experience actually ended up being at a car racing simulation at ViewSonic.

As the conference came to an end, we found ourselves very content with the progress we had made. We collected countless names and emails in our little notebook, and made connections and friendships that we’d otherwise not had the opportunity to make. At this time, I should say it. Thank you so, so much to everyone who stopped by our booth. We truly thank you from the bottom of our hearts. If you’re still reading up to this point, the one thing you might have noticed is that the throughline across this entire story is the kindness were were met with at every single turn. From the the ISTE employee who reprinted our badge without a second thought, to the Target staff smiling at our scheme, to (and especially) our booth neighbors sharing their power and kindness, everyone we met this trip was extremely good to us. If there is one thing I have learned from all this, it is that good karma exists. When you put goodness into the world, it has a funny way of getting back to you. And we were most definitely the recipient of that on our trip.

We look forward to using ISTE 2026 as momentum into our next phase of Deskpad. But for now, we sit back with content and gratitude, as we look back on what was truly, a hell of a trip.

- Akshay

At our core, we believe Deskpad is a mission. We are a group of high school students, and are concerned about AI’s impact on education. As profit-motivated AI labs push out better and more sophisticated frontier models, the industries that are slow to adapt have been the ones impacted most deeply. Education, which is built on the core values of transparency and accountability, has been hit particularly hard by AI’s continuous evolution. We have witnessed firsthand how our fellow classmates have explicitly taken assignments, uploaded them into ChatGPT, and copied-and-pasted the output back into their documents. This cheating has gotten so bad that schools across the country have had no choice but to ban the use of AI and technology altogether. While I once looked at many of these educators in disdain, after speaking to more than one hundred in my development of Deskpad, I have come to develop a deepening sense of sympathy for the seemingly insurmountable challenges they are currently facing in their classrooms.

What we created Deskpad to do, fundamentally, was to provide an alternative to schools who felt the need to ban AI in the classroom. Deskpad is not anti-AI. We are techno-optimist—of the belief that AI has the potential to do tremendous good for our society, and today’s students will be the ones pioneering that change. But all of these benefits will only exist if AI is adopted, and more importantly, adopted in the right way. Schools, to this point, have taken the wrong approach to AI. We believe that the solution to AI’s incorporation education is a guardrailed, transparent approach with both students’ and teachers’ best interests kept at heart.

The risk with AI in education, of course, is also especially large. The technology that allows a student to learn at a pace never seen before also has the potential to make them incredibly stupid. There is published research that an over-reliance on AI, especially in an academic setting, can erode the part of your brain that thinks critically. If AI is to be adopted in education, there are also immense concerns around AI’s environmental impact, student data privacy, and model hallucination on important facts. We believe, however, that several of these risks are mitigated when thoughtful and student-centered adoption is a conscious choice by schools. I will address our perspective on each of the previous concerns below:

1. On cognitive decline: If every AI model incorporated into the classroom is transparent and guardrailed, this becomes a significantly lower risk. Students who use AI as a “thought partner” are freed to produce their best work while still owning all of the content and ideas. While there will still be a need for some level of student surveillance (as there has been across education’s history), in a world where AI’s usage is allowed thoughtfully, student cheating will decline. The core idea is that when the technology is encouraged with the student’s best interests at heart, students will inherently begin to understand where AI will help or hurt them. In a world where the technology is flat-out prohibited, students will look for loopholes and workarounds, which is where the cognitive offloading truly persists.

2. On AI’s environmental impact: Of all the concerns we have studied, this is the one we are least comfortable waving away, because the costs are real. Training and running large models consumes meaningful amounts of energy and water, and we do not think it is honest to pretend otherwise. What gives us cautious optimism is the shape of the problem. The vast majority of AI’s energy footprint comes from training these models in the first place, a cost that is paid once by the labs and then spread across billions of uses, which means that a single student asking a question consumes a very small fraction of the total. We also believe efficiency is moving in the right direction, both because smaller and more specialized models are increasingly capable of doing what once required enormous ones and because the shift toward locally-hosted models tends to favor leaner systems that can run on ordinary hardware. None of this makes the impact zero, and we want to be clear that it does not. But we think the responsible path for schools is not to treat AI as uniquely wasteful and ban it, while leaving every other energy-intensive technology in the building untouched, but rather to adopt it deliberately, with an honest accounting of its costs and a preference for the most efficient tools that meet a student’s actual needs. This is the standard we hold ourselves to as well.

3. On student data privacy: While at this point effectively all student messages are being sent to a cloud provider (such as OpenAI, Anthropic, or Google), the world is moving towards locally-hosted models. This means that the LLMs that receive inputted messages run on the messenger’s own hardware, such as their computer. There has been promising research done in this space, and Apple themselves have taken the position that AI will run on user-owned hardware in the near future. For education, this means that while we currently have to trust cloud providers when they say they are not misusing student data, we will likely soon live in a world where this is not a concern at all. Furthermore, even current cloud providers are held to extremely rigorous data privacy standards, and there are massive penalties for any corporations found liable for misusing data, especially in the case of students and minors.

4. On model hallucination: Model hallucination has and will continue to be one of the most challenging issues that AI will face in its adoption. The encouraging news is that the risk is both shrinking and, more importantly, controllable. Newer frontier models hallucinate considerably less than their early predecessors, and peer-reviewed research has shown that the most effective safeguard is, somewhat counterintuitively, not waiting for a perfect model but changing how the model itself is used. When an AI system is grounded in a trusted, curated body of source material, such as a syllabus or assignment, (rather than answering from memory) and is permitted to admit when it doesn’t know something, hallucination rates drop dramatically, in some studies to near zero. We have taken this exact philosophy with our chatbot Sage which exists on Deskpad’s platform Learn. By anchoring responses to vetted educational material and designing for transparency over confident guessing, we can give schools the benefits of AI while keeping the factual reliability that a classroom demands.

We believe that if all of these concerns are addressed honestly and diligently, AI truly can change the way people learn for the better. If we live in a world where education and information is truly democratized, our populations will be smarter, happier, and more likely to create positive change. In every instance in human history where education has been made a top priority, the population has thrived. We believe this moment is no different. Our population can either massively benefit, or massively suffer, from the adoption of AI. If we want to benefit, we must band together as a population, and adopt it on our own terms. This is the sole reason for Deskpad’s existence, and why we believe it is such an important problem to be solving.

Across the country, students, as they tend to do, began immediately exploiting this new tool in the classroom, and before long, teachers were caught in a frenzy of AI-generated submissions that they could not tell apart from genuine student work. The first solution to this problem made its way to the public sphere in January of 2023, just a month and a half after ChatGPT’s initial launch. Princeton undergraduate student Edward Tian launched the AI-detection tool GPTZero, which claimed to “use ChatGPT to detect itself.” Tian had been deeply immersed in natural language processing (NLP) research at Princeton, and his proposed solution seemed like a logical next step for many educators who had gotten used to the Turnitin “gotcha”-style approach to student integrity. For those unfamiliar, Turnitin’s claim to fame that drove them to an estimated $150M+ annualized revenue and an acquisition price of $1.75B in 2019 was the Similarity Score. John M. Barrie, along with collaborators Christian Storm, Emmanuel Briand, and Melissa Lipscom, founded Turnitin in 1998 as PhD candidates at UC Berkeley. Their philosophy was straightforward. Academic cheating—especially plagiarism—had seen a tremendous uptick with the growing adoption of the Internet. Students now had a seemingly infinite list of sources to pull writing from (and pass it off as their own), and educators everywhere were struggling to deduce what work was authentic and what was copied and pasted in from an external source. While some ignored the problem, many were actively searching for a solution, which is exactly where Turnitin came in. The Similarity Score was an algorithm that was designed to compare a student’s work to existing published work on the internet, and generate an integer percentage between 1 and 100 on how likely the work was plagiarized. The solution felt deeply ahead of its time, and for the decades to come, its adoption became widespread in nearly every academic setting.

GPTZero promised educators something familiar. Just as the Similarity Score gave teachers a percentage for plagiarism, GPTZero and the wave of detection tools that followed offered a percentage for AI-generated content. Run the student’s work through the system, get a number back, make a judgment call. For teachers who had spent years relying on Turnitin as their safety net, AI detection felt like the obvious next chapter.

But there was a fundamental problem with this approach that most people didn’t see at first. Turnitin’s Similarity Score works because plagiarism is a matching problem. You’re comparing one piece of text against a database of existing texts and looking for overlap. Either the words match or they don’t. It’s not perfect, but the underlying logic is fairly sound. AI detection is a completely different kind of problem. You’re not comparing text against a source. You’re trying to determine whether the text itself was produced by a human brain or a language model, and you’re doing this by looking at statistical patterns like perplexity and burstiness. Perplexity measures how predictable each word in a sentence is, and burstiness measures how much variation there is in sentence length and structure. The theory is that AI-generated text tends to be more uniform and predictable, while human writing is messier and more varied.

The issue is that these patterns are not reliable indicators of authorship. A student who writes clearly and concisely can produce text with low perplexity that looks identical to AI output. A student who learned English as a second language might write with patterns that a detector interprets as machine-generated. And a student who happens to organize their essay in a structured, methodical way will trigger the same flags that ChatGPT does. In 2023, professors at Texas A&M University accused an entire class of using AI on their final assignments based on detector results. Some students were threatened with failing grades or holds on their diplomas. Many of them had not used AI at all.

Then there’s the other side of the arms race. Within months of GPTZero’s launch, tools like Quillbot, Grammarly Humanizer, and countless others emerged that could take AI-generated text and “humanize” it, introducing enough randomness to fool detectors. The cost of evading detection dropped to zero almost immediately. The students who were actually cheating found workarounds in minutes, while the students who wrote honestly were the ones getting flagged. The system was punishing the wrong people.

What we’re left with is a landscape where schools have been forced to choose between two broken options. Some have banned AI entirely, blocking ChatGPT on school networks and treating any use of it as academic dishonesty. Others have gone the opposite direction, allowing unrestricted use with no guardrails and no visibility into what students are actually doing. Neither approach works, and both are trying to answer the same question: “Did this student use AI?” I think that’s the wrong question. It treats AI use as a light switch, on or off, when in reality there’s a massive spectrum between a student who pastes “write my essay” into ChatGPT and a student who uses AI to pressure-test their own argument before revising. Both of them “used AI.” Only one of them learned anything. The question education should be asking is not whether students are using AI. It’s whether they’re thinking. And right now, nobody has a good way to answer that.