AI today feels a bit like air conditioning—quietly humming in the background, making life easier for those who use it, while others ignore it or resist turning it on. But like any great innovation, it casts its shadow far ahead, hinting at profound changes to come. The question is not whether AI will transform us, but how—and whether we’re ready for what it asks of us and what it might quietly take away.
What excites me, though, is not the spectacle of AI’s potential, but the quieter, more profound questions it forces us to ask about ourselves. At its core, AI doesn’t create new knowledge—it identifies patterns in text, images, and data, patterns that we humans wrote and shaped. It reflects us back to ourselves, for better or worse.
And that’s where the real question lies: Which patterns do we want to perpetuate, and which should we let go? What should be handed off to machines, and what should remain uniquely human? It reminds me of the Cherokee tale of the boy who dreamed of two wolves battling in his dreams. One wolf was angry, vengeful, and destructive, while the other was kind, generous, and peaceful. Troubled, the boy asked his grandfather which wolf would win the fight. The old man replied simply: The one you feed.
In many ways, this is where we stand with AI today. It reflects back to us not only the best of what humanity has written, but also its biases, flaws, and darker impulses. The question, then, is not just whether AI will change us, but which parts of ourselves we choose to feed as we teach and shape it. What will these transformations demand from us, and what might they quietly take away? Let’s consider five key areas where these shifts—both exciting and unsettling—are poised to redefine how we work, live, and interact.
1. The rise of voice: can we talk?
There’s something primal about the act of speaking. Voice is our first technology, the original interface. In a way, AI bringing us back to our voices—after decades of typing, tapping, and swiping—feels like a homecoming.
Already, tools like ChatGPT, Gemini, and the endlessly awkward Siri have started turning voice into a viable interface. I’ve found myself asking these systems for everything from driving directions to restaurant recommendations, and though I still sometimes feel silly talking to a machine, the utility of it is undeniable.
The thing about voice, though, is its intimacy. When you ask an AI assistant for help, it doesn’t feel like inputting a command—it feels like conversation. I imagine this could be transformative in certain situations. Teachers could create lesson plans aloud, while elderly individuals might dictate memoirs to an AI that organizes their thoughts. Travelers could point their phones at murals and have AI narrate their histories in real time, in the language of their companions.
But voice is also personal, revealing more than just words. It carries tone, emotion, and context—things we may not always want to share with a machine. And yet, as voice AI improves, it may begin to feel like a natural extension of ourselves.
Erika Hall, Conversation Design specialist wrote “Conversational design is truly human-centered design, every step of the way. There is no next big thing, only the next step in an unfolding story of how people use technology to be more themselves.” Thank you Lily Chambers, my co-editor for this addition.
Prediction: By the end of 2025, voice AI will become second nature for most of us—just another way to engage with the digital world. It will be mundane, yes, but also intimate, and therein lies its charm and its complication. Look around and you’ll see young teens using AI as companions. Where does this lead? It’s a question I’m asking myself almost daily.
2. Natural language coding: the end of gatekeeping?
I have a confession: I’ve always admired coders, though not without a hint of resentment. The mystique of their craft—the way they summon something out of nothing—has always felt inaccessible to me, like a modern kind of alchemy. But AI is dissolving that mystique, and I’m not sure how I feel about it.
Tools like GitHub Copilot and Anthropic’s coding assistants already allow developers to write and debug code at lightning speed. But the real game-changer is this: AI is bringing coding to the non-coders. Imagine telling an AI, “Build me a website where users can share recipes and rate them,” and watching it materialize. This democratization of technology is thrilling, even if it feels like we’re outsourcing creativity to an efficient but impersonal partner.
And yet, I wonder: Will the elimination of technical barriers lead to an explosion of creativity, or simply an abundance of mediocrity? There’s something to be said for the friction of learning a skill, for the patience it requires. When tools are too easy, do we lose some of the art?
Prediction: By 2025, natural language coding tools will empower teachers, designers, and entrepreneurs to build digital experiences without needing technical expertise. Coding, as a skill, will no longer be the gatekeeper it once was. Whether that leads to greater innovation—or a glut of half-baked projects—remains to be seen.
3. AI-generated video: the meme renaissance
Video feels like the final frontier for AI. Text can inform, images can provoke, but video demands something deeper—a narrative, a rhythm, a mood. In 2024, AI video tools like Runway, Veo, and OpenAI’s Sora turned this medium into a playground for creators, and the results were both dazzling and unsettling.
The possibilities are intoxicating: marketers creating entire ad campaigns in minutes, teachers producing engaging tutorials, even everyday users generating videos that rival Hollywood effects. But the reality might be more prosaic. Most of us will use these tools to make memes and influencer content.
I can’t help but feel both excitement and exhaustion at the thought of an internet flooded with AI-generated videos. On the one hand, this democratization of creativity is thrilling. On the other, I wonder if the sheer volume of content will dull our senses. Will we still marvel at a beautifully crafted video when we know it was created in seconds by an algorithm?
Prediction: By 2025, AI video tools will saturate every corner of the internet. From TikTok to advertising, video content will be easier to create than ever—and, paradoxically, harder to impress with. But coupled with AI voice, will people begin creating their own imaginary worlds to live in? I wonder.
4. Small models, big impact
Not all AI needs to be a towering genius. Sometimes, the simplest tools are the most transformative. Small Language Models (SLMs)—the lightweight cousins of large models like ChatGPT—are already proving their worth in everything from text prediction to smart home devices.
What’s remarkable about SLMs is their specificity. Unlike massive, general-purpose AIs, these models can be trained on a company’s proprietary data or optimized for niche tasks. They’re faster, more private, and far less resource-intensive. In short, they’re practical—a word not often associated with AI hype.
Prediction: By the end of 2025, SLMs will be everywhere—powering tools you didn’t even know were AI-driven. They will be the steady, reliable workhorses of the AI world, and we’ll wonder how we ever lived without them. Running AI tools locally brings back a feeling of privacy and security and I’m all for that.
5. The year of the AI agent
Here’s the thing about AI agents: They’re both exhilarating and slightly unnerving. These are systems designed not just to answer questions or generate content, but to act autonomously on our behalf. This is where people always promise some kind of magic agent to manage your calendar and book your travel. I’d let that idea go for now.
A better way to think about these as little semi-autonomous programs that do things for you and bring back a result. Like download the latest proposed state or federal legislation and analyze it from the perspective of my needs and tell me if I should write my representative or senator. That’s what I want.
Prediction: By 2025, businesses will integrate AI agents into their operations, using them to automate workflows and experiment with new processes. But for consumers, these agents will remain just beyond reach—too fragile, too untrustworthy, too unsettling for most people to welcome into their personal lives.
Closing reflections: progress and Its shadows
As we feed the good wolf—the patterns we want AI to reflect—another truth emerges: AI forces us to confront our ambivalence about progress. We crave efficiency, yet mourn the loss of effort. We embrace convenience, even as it erodes our patience.
In 2025, AI will quietly insert itself into our lives, becoming both indispensable and invisible. And maybe that’s the best outcome: a technology that doesn’t dazzle or terrify, but simply works. Progress, after all, doesn’t always arrive with fanfare. Sometimes, it just hums quietly in the background, like a shadow cast by something still taking shape. So when suddenly you realize you have some extra time, why not just sit quietly and let yourself be in the moment.
Share this post