When ChatGPT first arrived, it felt like we'd jumped a couple of decades ahead in time. It was unreal, and a little threatening. The media filled with questions about what it all meant, the loudest being the obvious one: is AI going to take our jobs?
The fear was real, and many of us judged generative AI without ever exploring it. Almost a year on, we have a whole landscape of tools built on these models: writing assistants, tutor chatbots, text-to-image generators, lesson-plan builders, and more.
Over that year I've used AI in plenty of ways, first as a learning designer and then as a student here at HGSE, and I can safely say I've gotten over the fear of it taking my job. But in recent conversations with peers and mentors, I learned there are still real misconceptions, especially across K-12 and higher education, that have led some institutions to ban these tools outright.
AI is here to stay, and to grow in ways we can't yet imagine.
I share the concerns about data safety, privacy, and ethics. And it's also worth accepting that AI isn't going anywhere. To prepare for that future, policymakers, leaders, and educators across every industry need to understand how generative AI actually works. So how do we get there? Here are a few takeaways and best practices I've gathered, especially for working with large language models.
1. Augment, don't outsource
When ChatGPT launched, I treated it as a tool built to do my job for me. As I played with it, I realised that while it's a great place to start, it doesn't always produce accurate or reliable information. The real skill is partnering with it: using it to extend your capabilities and augmenting its output with your own judgement and expertise, rather than shipping whatever it generates.
2. Build strong GenAI policies
To make that kind of use possible, you first have to create awareness about what AI is, how it can be used, and what resources exist. At HGSE, every course has an AI policy that spells out the encouraged ways of using it for coursework. The university also built a sandbox to explore these tools more safely, plus workshops, user guides, discussion spaces, and demos for students and faculty alike. That fosters a healthy, shared way of learning together.
3. Explore, and stay curious
In my experience, the best way to learn generative AI is to play. Have fun with the questioning and prompting in all sorts of ways. Use it to write a recipe, sharpen your writing, or understand a tricky concept. With custom agent builders, it's getting more playful even for those of us who don't code. The internet is full of guidance on how to use AI well, and how not to. Starting there and experimenting is a great way in.
As I said at the start, the one thing AI can't do yet is make judgements the way humans do. It's on us to use these tools creatively and ethically, to make ourselves more capable.