Microsoft bets on friendlier AI companion called Mico
Add Axios as your preferred source to
see more of our stories on Google.

Illustration: Sarah Grillo/Axios
Microsoft rolled out a new AI companion Thursday called Mico, designed to be more personal, supportive and what the company calls "human-centered."
Why it matters: Microsoft's early partnership with OpenAI kept Copilot focused on productivity, leaving the friendliness to ChatGPT. This update blurs that line.
Catch up quick: Microsoft and OpenAI entered into an exclusive computing partnership in 2019, and since then the Copilot maker has invested billions in the ChatGPT maker.
- But the situationship has grown complicated — and at times strained.
- OpenAI's efforts to become a for-profit endeavor have created issues between the two, and the product overlap is confusing.
The big picture: Mustafa Suleyman, who co-founded Google DeepMind and now leads Microsoft AI, announced Mico in a blog post Thursday.
- Suleyman said the company designed the new Copilot to help humans make better decisions, not to replace human judgement.
- "I often say you should judge an AI by how much it elevates human potential, not just by its own smarts," Suleyman said.
- Microsoft says Mico "challenges assumptions with care" and "adapts to your vibe."
The intrigue: Suleyman also co-founded Inflection, a startup claiming to launch the first "emotionally intelligent chatbot," called Pi.
- Pi launched around the same time as ChatGPT, but withered on the vine when Microsoft acquired most of Inflection's workforce, including Suleyman, in March 2024.
- In a personal blog post in August, Suleyman outlined what he meant by a personable chatbot. "We must build AI for people; not to be a digital person," he wrote.
Zoom in: Microsoft is also betting that people want AI to be a social experience, although so far that has not proven to be true.
- The Groups tool lets users brainstorm, co-write and work together with up to 32 people ... and Mico.
Between the lines: Microsoft won the workplace early in terms of productivity software, but the company also has a long history of releasing human-centered tools that missed the mark.
- Microsoft Bob was a mid-1990s attempt at a more friendly user interface for Windows, one so roundly mocked and criticized that it was discontinued after a year.
- Clippy, the late 90's assistant in Microsoft Office, nagged workers from the corner of their desktop. It was hated, but sometimes heaps of scorn can mean a kind of affection. Using the Clippy icon as your profile picture has become a silent protest to resist the use of AI that violates copyright.
- The "Me" in Windows Me stood for Millennium Edition, but Microsoft also cheekily meant the "Me" to stand for its primary target — home users. Released to the public in 2000, the OS was rife with bugs and was widely considered a failure.
- Microsoft was one of the first to friendly chatbots with the 2016 release of Tay, which quickly became a lesson in the need for serious AI safety training, especially when releasing a bot onto social media.
Go deeper: AI's mental health fix: Stop pretending it's human
