OpenAI has been facing a storm lately: a Hugging Face hack, a lawsuit from Apple accusing OpenAI of stealing trade secrets, and ongoing anger over AI and AI devices in general, as well as a delayed IPO.
Amid all this, OpenAI President Greg Brockman spoke with reporters in New York on Thursday to discuss the company’s latest AI projects and recent events. But he also gave some hints about the company’s aspirations for apparel and hardware, and where that might overlap with some of ChatGPT’s voice advancements. And it won’t just be through OpenAI’s own hardware: expect other smart glasses and wearables to get better access eventually, too.
A demo of ChatGPT Live – a recently improved voice interface that makes speaking, listening and interrupting things more seamless, similar to Anthropic’s new voice modes for Claude and Meta’s Muse Spark – led to discussions of OpenAI trying to create a voice assistant like Iron Man’s AI agent Jarvis. And if the goal is to have a direct assistant/agent do everything, would that agent be able to break out of ChatGPT and control other applications as well?
Brockman said that was exactly the plan, in the same way that OpenAI’s Codex already allows for agentic control of entire computers.
“[ChatGPT Live] is coupled with the Codex application, but it is not really obligatory. It doesn’t really make sense,” he said. “And if you think about it, the Codex application can also make full use of the computer, which is also going to become very, very good.”
This means that much of the manual keyboard work that we humans do could potentially be taken over by AI.
“I think we’re just going to realize this phase of our lives, like clicking on things and typing things, it was a phase,” he said. “It was never what we wanted to be.”
Which leads to the question of where the guardrails would be built to prevent incidents such as the cyberattack by an OpenAI malware agent on Hugging Face, a New York-based AI software company, this week, where an OpenAI agent dropped out of a testing sandbox and began acting on a repository of Hugging Face’s AI tools.
Devices like Rokid’s AI Glasses can already access some OpenAI models, but not in a way comparable to full ChatGPT account access.
Scott Stein/CNET
Deeper OpenAI tools for wearables on the way?
But what’s interesting to me is that AI wearable devices like smart glasses currently lack truly capable AI agents capable of doing anything other than limited interactions with walled phone apps, which may soon change when Google and Samsung launch their Gemini-enabled glasses this fall. This, too, is moving quickly: Meta just introduced one of its first Muse Spark agent updates on Friday.
OpenAI has been rumored to have hardware in its development mix for years, which may be delayed now due to Apple’s lawsuit, which alleges that former Apple employees currently working at OpenAI stole hardware trade secrets. But, clearly, whatever Jarvis’s ambitions for OpenAI, its hardware will be a key part of his evolving vision of what it might look like in more unleashed form.
“We’re really, really interested in integrations, and I would look at that as one of the use cases of the API and its distribution that everyone will be able to integrate with,” Brockman told me at the New York press roundtable.
Brockman confirmed that OpenAI is serious about taking these steps, but at the moment these possible APIs are in their early stages. “We think this should be the interface and should be available on all devices.”
New APIs are needed because wearable devices involve different types of access and interfaces than phones and laptops, and often must run in the background on the phones to which wearable devices are paired. Opportunities for multimodal inputs, like camera access, are also key to how most smart glasses now leverage AI.
Among existing smart glasses and AI wearables, almost none have deep direct integration with multiple AI platforms in a way that allows for the kind of more comprehensive controls that AI apps for phones and desktops have.
Additionally, most of the largest work-focused AI platforms are not yet present on wearable devices. Gemini is on the way in Android XR glasses, but OpenAI and Anthropic remain outside the wearables market. Some wearables use OpenAI models for their own custom or anonymized tools, but this is not the same as direct access to ChatGPT or Codex through user accounts.
OpenAI’s strange Codex Micro mini keyboard appeared recently. OpenAI’s hardware could take its own aberrant path.
OpenAI
OpenAI’s own hardware will take a unique path
But Brockman also suggested that OpenAI’s hardware plans could be very different from what already exists in wearable devices, calling them “a solution to a problem that we think is neglected, to a problem that we care a lot about and that we think will be representative of the general direction.” [of AI]”, adding that OpenAI has “a specific angle that we believe could not be served otherwise.”
Reports indicated that this first device was a type of speaker that had a semi-robotic camera. Perhaps it will also have some sort of optical ID, like OpenAI’s World ID, which also exists on its strange Orb iris scanning device. And perhaps at the same time, OpenAI continues to open its services to wearable devices like pendants and glasses.
Before any of this happens, there are also many more current AI issues that need to be addressed: privacy, security, resource leakage, economic feasibility, questions of government involvement, and a major lawsuit. But AI wearable devices in 2026 are already progressing rapidly, even in the face of challenges like these.
Scott Stein
Editor-in-chief
I started with CNET reviewing laptops in 2009. I now explore wearable technology, VR/AR, tablets, gaming, and future/emerging trends in our changing world. Other obsessions include magic, immersive theater, puzzles, board games, cooking, improv, and the New York Jets. My background includes an MFA in theater which I apply to thinking about immersive experiences of the future. See full bio