As artificial intelligence (AI) continues to transform industries, the importance of cybersecurity has evolved from a reactive necessity to a proactive strategy. In sectors like mental health, education, and wellness, where sensitive data and vulnerable user groups are involved, AI security is not just a technical requirement, it’s an ethical one. At curaJOY, where we build AI-powered tools that support children’s emotional and behavioral development, we understand that user trust is foundational. That trust can only be preserved through a defensive, forward-thinking approach to both cybersecurity and AI integrity. A proactive, defensive approach means anticipating and mitigating threats before they materialize. While traditional cybersecurity models often rely on detecting and responding to attacks, AI introduces new complexities: data poisoning, model manipulation, adversarial inputs, and privacy breaches stemming from model training itself. These risks aren’t theoretical — they happen in real time across industries. For organizations like ours, the responsibility to safeguard against such vulnerabilities goes beyond compliance; it’s core to our mission of empowering families and protecting the most sensitive information they share. What Proactive, Defensive Security Looks Like (and What We’ve Done at curaJOY) 1. Secure by Design — Not as an Afterthought What curaJOY does: From the start, we bake in threat modeling during development, especially focusing on how attackers might exploit the AI models or manipulate the data that trains them.Common mistake: Teams often treat security like a “final checklist” before launch instead of a continuous process. 2. AI Model Monitoring and Drift Detection What curaJOY does: We deploy real-time monitoring to flag suspicious outputs from our models. If an AI system suddenly behaves differently, say a therapeutic chatbot giving out-of-scope responses, we investigate and retrain.Common mistake: Many startups “set and forget” their AI models after deployment, leaving them vulnerable to data drift or subtle adversarial behavior. 3. Data Privacy as a First-Class Citizen What curaJOY does: We anonymize, encrypt, and regularly audit user data. We’ve also limited model training to exclude personally identifiable information (PII) whenever possible, to reduce risk in case of leaks.Common mistake: Collecting or using more data than needed, especially sensitive personal data, without clear boundaries or user control. 4. Zero Trust and Role-Based Access What curaJOY does: Internally, we follow the Zero Trust model. Access to production models and data is strictly role-based and continuously reviewed.Common mistake: Over-permissioned systems where too many team members have access to sensitive environments. We’re also keeping a close eye on emerging AI security challenges. These include prompt injection attacks on language models, model theft through repeated querying, and bias exploitation, where users intentionally provoke offensive or flawed outputs. By actively participating in the broader AI security and research community, we stay informed and continuously refine our safeguards. This commitment allows us to protect our models and users from today’s threats and tomorrow’s unknowns.To help others navigate similar challenges, we’ve identified a few common pitfalls to avoid: skipping security planning in AI development, trusting AI outputs blindly, neglecting post-deployment monitoring, excluding security experts from product planning, and assuming open-source tools are secure by default. Avoiding these missteps can dramatically reduce your exposure to both technical and reputational risk.Ultimately, at curaJOY, we believe that security is an act of empathy. It’s about valuing users’ trust enough to protect it at all costs. It’s about recognizing that AI isn’t just software — it’s shaping conversations, decisions, and even emotions. In this landscape, security isn’t a backend issue. It’s a user experience imperative, a brand differentiator, and above all, a responsibility we carry with pride.

Our booth at the AI Show @ GSV + ASU was tiny—just 3×3. We were flanked by past winners of the Tools Competition, NewSchools Venture Fund, and Fast Forward Accelerator—programs we’re now going through ourselves. It was inspiring to be among them—but I’ll admit, I had a moment of doubt. Had I waited too long? Most startup advice says the founder’s job is to fundraise, first and foremost. But I chose to build first, not sell or pitch. Down the aisle were larger, more established companies with massive booths and polished demos. And yet, I found myself quietly comparing: while others were showing pieces of the puzzle, we’ve spent three years building an end-to-end solution—designed with and for the people it serves. It wasn’t a judgment. Just a realization: how different curaJOY is and how substance can grow from small spaces. In the AI4C Accelerator—sponsored by Google.org and AWS—we’ve found ourselves on the more mature end of the spectrum, with an AI strategy and infrastructure already operating in real-world settings. Every organization in the cohort brings something unique to the table, and for curaJOY, our edge came from years of quiet groundwork: building responsibly, listening to the people we serve, and resisting the pressure to launch before we were truly ready. Still, I couldn’t shake the guilt–We could’ve gone to market earlier. We had the talent. We had the science. We had the need. So why didn’t we? Because sometimes, going fast isn’t going forward. We Didn’t Just Build an App—We Built an Ecosystem curaJOY spent three years in deep product development. Coming from the electronics industry, where product cycles rarely last more than a year, I used to feel uneasy about that number—three years. But this wasn’t consumer tech. We weren’t optimizing for speed; we were safeguarding lives. Unlike traditional tech rollouts, we were operating in a space where the stakes are deeply human. When a platform claims to support a child’s behavioral or emotional development, the margin for error is razor-thin. Getting it wrong doesn’t just lose a user. It can do harm. We weren’t just cautious—we really care, so we were concerned: that a rush to market would teach families to self-diagnose and self-treat without oversight; that irreplaceable professionals—therapists, teachers, clinicians—would be undermined or replaced; that we’d contribute to the very burnout and misinformation we set out to solve. That’s why curaJOY didn’t rush. We trained clinicians, educators, youth, musicians, and parents in AI/ML and UX. We built safety nets with human escalation triggers. We ensured cultural adaptation went beyond translation—because behavioral expectations, family roles, and even reward systems vary widely across communities. And we made clinician supervision—not optional, but foundational. This is why the tech giants haven’t cracked behavioral health. And it’s why venture-backed autism centers often collapse: faster and bigger is not always better in this space. The Hidden Cost of “Move Fast and Break Things” Big tech is built to optimize—for clicks, retention, and revenue. But behavioral health isn’t about quick wins. You can’t A/B test which way lashes out less trauma. You can’t “growth-hack” your way into trust between a therapist and a child.We’ve already made this mistake once. We let social media and constant connectivity reshape childhood without a plan—and now we’re paying the price in rising loneliness, shrinking attention spans, and mental health crises. It’s not that these technologies were inherently bad. It’s that we let them lead, instead of leading them. With AI, we have a second chance. Let’s do things differently this time. Let’s not go blind into this one. We’ve seen what happens when technology rolls out without foresight:– Social media reshaped childhood and relationships without a wellness blueprint– Apps promised “mental health support” but lacked supervision, equity, and follow-through– The behavioral health workforce is burning out—and burning out silently curaJOY made the difficult decision to go slow because the easy path would have cost us integrity—and maybe even lives. Ready for Prime Time, and Built for the Long Game Today, curaJOY is ready for scale—not because we went fast, but because we built right. Our AI doesn’t just chat—it collaborates with teachers, clinicians, and families to surface root causes of behavior. It drafts FBAs and support plans. It nudges users toward completion. It logs every step for transparency and compliance. And it delivers culturally responsive, 24/7 behavioral support in the places kids actually turn to—like Discord, WhatsApp, and social media. We’ve built partnerships with school districts, outpatient clinics, and global researchers. We’ve validated our models with PhD-level clinicians. And we’ve co-created our system with youth ambassadors, educators, and caregivers.We’re not just launching a product. We’re shaping how AI will enter education and healthcare—safely, ethically, and equitably. Going Places—Intentionally Yes, I wish our incredible team saw the fruits of their labor sooner. But I’m proud that we didn’t trade speed for responsibility. The part that weighs on me most isn’t lost time—it’s the people. The volunteers who’ve given their weekends, the clinicians and engineers who’ve built beside me without pay, trusting the mission. I haven’t paid myself in years, and I fund curaJOY personally—but not everyone can do that. I owe it to them to scale this the right way. To ensure their sacrifices turn into impact—and into sustainability. curaJOY is a grassroots nonprofit by design, because lasting solutions to access and equity challenges must be shaped by the communities they serve. Real change takes time—and that’s something we seem to forget.Some of society’s biggest problems—like the behavioral health crisis or inequities in education—didn’t appear overnight. Yet we keep falling for the illusion of overnight miracle cures. It’s like expecting to undo years of weight gain in a 30-day crash diet. We know that doesn’t work for our bodies—so why do we expect it to work for systems this complex? We can’t patch over long-standing pain with quick fixes. We need to build intentionally. Inclusively. Responsibly.curaJOY is ready to scale not because we sprinted, but because we prepared. If you’re a funder, educator, technologist, or healthcare leader who believes in getting it right—and just getting it done—let’s talk. Good things take time–It’s easy to forget this in our society of instant fame, viral launches, and one-click everything. But responsible technology and thoughtful innovation can’t be rushed—especially in healthcare and education. We’ve seen what happens when tech moves too fast and too big. Innovation with integrity cannot be rushed. It’s designed, tested, and built to last.

A few days ago was my first day at the AI-GSU show. The experience was….different than I was expecting to say the least. The building was bigger and was less packed with people than I thought it would be. There was a machine making cotton candy for free, along with some people giving out coffee. Also, a lot of the tables were smaller than I thought they would be. “That tiny space costs so much?” I wondered. “What made renting out that table so worth it?” As I would later find out, the table is just the set up, it’s the people and the connections they make that matters. As a youth ambassador, I was tasked with going out to collect interviews for the research project tapestry of emotions. What I had to do seemed relatively simple: Get someone to talk about one specific emotion and how it feels for him or her (Ex: A time you felt this, physical sensations associated with the emotion, what is the most common emotion for an age demographic, etc). However, there was one major hurdle to getting the interview: Actually getting a person to agree to do it. People will say no if given any reason too, and even I knew that simply going up and asking someone “Can you take this survey for a research project I’m doing for a nonprofit?” would almost guarantee a “No”. To help solve this, Caitlyn, the CEO and founder of curaJOY, gave me some advice: Talk to the person first, make him or her feel on top of the world, like you’re coming to him or her specifically for advice. Then segway into getting the interview. With that knowledge in mind, my group and I set off. But as it would turn out, I would need a lot more help than just that. My first problem was actually going up to talk to people. As an introvert, I don’t like large crowds. Or going up to someone and starting a conversation. But that was what I was assigned to do, so I somehow managed to go and start talking to people. There I ran into my second problem: How to steer the conversation towards the tapestry of emotions research project. I was told to let the people talk about themselves first so that is what I did. The problem was, I was taking a too roundabout method, trying to wait for an opportunity to ask “Would you be willing to answer a question for my research project, tapestry of emotions?” I wasn’t actively trying to seize my moment, I was lying in wait. So no wonder that my using the roundabout method failed. Either the conversation dragged on too long or by the time I had the nerve to bring up curaJOY, the pivot seemed too unnatural. Frustrated, and only after ten minutes at that, I walked back to curaJOY’s booth. I told the CEO that I was not cut out for this, that I couldn’t do this, and that she should just stick me in filing for curaJOY’s stuff, where I wouldn’t have to talk to people. It was then the CEO realized that I needed some more guidance. Leaving the booth in the hands of another curaJOY member, she showed me how it was done. She went up to a table and started talking to the people. I didn’t know how, but next thing I knew, she had gotten an interview. It was at this point I decided to pay extra close attention to figure out the process. Overtime, I began to notice a general pattern. The method the CEO used was the same one she instructed me with, but somehow, when shadowing her, I understood it even more. Just having the information on how to do something isn’t always enough; sometimes, there has to be a live demonstration. That is why teachers do a demo before letting kids try something on their own. If it can work in the classroom, why can’t it work in a business environment? Shadowing the CEO, watching her do a few demonstrations and then trying it out myself was the secret formula for success I had been looking for all along. Somehow, I managed to get an interview from almost everybody I tried talking to. The interviews were good as well! I set a record among the youth ambassadors, getting 9 interviews on my first day! My first day, the day where I told the CEO that I couldn’t do it, but I did! I did not break my record the next two days, but I still managed to bring in a large number of interviews, and the other youth ambassadors and I managed to meet our quota of 50 interviews. Words can’t express how amazed I am not just at my group’s progress, but also of my personal growth as well. The trade show gave me a taste of the business world, and now, I feel more prepared for it.