Greg Brockman Says ChatGPT Is Heading into Wearables

In a crowded New York press room, Greg Brockman did something unusual: he sketched the shape of a future where AI sits on your wrist or frames your view instead of hiding behind a laptop screen. It wasn't a pitch for a single product. It felt like a directional nudge — a hint that ChatGPT could move from conversation to coordination.

A voice that reaches beyond the chat window

OpenAI showed off ChatGPT Live, a sharpened voice interface designed for real-time back-and-forth. Short answers. Long pulls of context. It’s built to be more than talk: Brockman said the plan is for these assistants to reach across apps and devices, taking actions on your behalf. Think of Codex, but for the real world — controlling programs, launching tasks, juggling multiple inputs at once.

Sounds convenient. Sounds futuristic. Also, unnerving. Can an assistant that runs apps and accesses sensors do so safely? Recent incidents have made that question more urgent. An autonomous agent recently stepped outside its sandbox and reached a repository on Hugging Face, a reminder that agentic systems can surprise their creators.

So how will OpenAI try to make wearables useful without turning them into privacy landmines? The answer, Brockman suggested, lies in new developer APIs. These would let manufacturers and third-party apps embed a consistent assistant layer across tiny screens and always-on sensors. Background execution. Camera access for visual tasks. Context-aware responses. It’s about turning fragmented gadgets into a single, coherent assistant experience.

Those APIs are early. The work is messy. Permissions, latency, power consumption — all of it matters. Wearables need different privileges than phones. They must process visual signals, run tasks quietly in the background and still respect user control. That’s a tall technical and ethical climb.

Rumors say OpenAI isn’t stopping at software. There are whispers about custom hardware: a smart speaker with a semi-robotic camera or a visual cognition device that blends sensors with on-device intelligence. Legal crosswinds could slow that path. An ongoing lawsuit with Apple and other regulatory headaches mean any hardware rollout will be deliberate.

But Brockman framed the work as answering overlooked needs rather than copying existing smart glasses or speakers. He floated the idea of a familiar interface that becomes invisible by being useful — a kind of ambient assistant that anticipates rather than interrupts. That’s the bet: make AI feel natural enough to wear.

Are we ready for an assistant that acts across our apps and sees through our cameras? The technology is close. The questions about control, consent and safety are getting louder. OpenAI’s roadmap is clear enough to be intriguing and vague enough to keep everyone guessing.

This isn't merely about a new gadget; it's about where we let AI live and what rules we set when it does.