OpenAI and designer Jony Ive are trying to solve a number of technical challenges before releasing their mysterious AI gadget, according to a Financial Times report. Two of the challenges include figuring out the palm-sized device’s “personality” and how often it should talk. “The concept is that you should have a friend who’s a computer who isn’t your weird AI girlfriend,” according to one source briefed on the plans.

The report describes the device as lacking a screen and being “roughly the size of a smartphone” that can be either carried around by the user or placed on surfaces like a table or desk, mirroring similar details that leaked in May. It’s designed to communicate with users through a microphone, speaker, and camera — or multiple cameras, according to one FT source. This is the first of a family of devices being developed by Ive and OpenAI CEO Sam Altman, targeted for release sometime in late 2026 or early 2027.

OpenAI is looking for ways to make the gadget “accessible but not intrusive,” according to the FT’s sources, with the aim of providing a similar experience to Apple’s digital voice assistant Siri, “but better.” The company is struggling to decide on the voice and mannerisms for the AI-powered device, and ensure that it knows when to engage with users and finish a conversation. One source in the report said OpenAI is taking an “always on” approach that gathers data throughout the day, rather than being activated by specific verbal prompts like Amazon’s Alexa-powered Echo devices.

Privacy issues around the always-listening device and budgeting constraints for computing infrastructure must also be overcome, according to the report. “Amazon has the compute for an Alexa, so does Google [for its Home device], but OpenAI is struggling to get enough compute for ChatGPT, let alone an AI device,” a source close to Ive told the FT. “They need to fix that first.”