"Make people addicted." That phrase appeared in internal Microsoft documents describing the company's strategy for its new AI assistant, according to leaked planning materials.
The language is striking because it's rarely what you see in official communications. Most companies talk about "engagement," "daily active users," or "habit formation." Addiction is the same concept with the pretense stripped away - and when it shows up in internal strategy documents, it signals that the engagement-at-all-costs playbook from social media may be influencing AI product design.
Microsoft has been building AI into nearly every product it sells. Copilot appears in Windows, Office, Teams, and Edge. The company invested roughly $13 billion in OpenAI and is reorienting its entire consumer and enterprise software stack around AI chat. Getting people to use these tools daily is central to the business case for all of that spending.
The concern with "addicted" as a design goal is that it tends to prioritize retention over usefulness. Products built around addiction metrics add friction when you try to stop, optimize notifications for engagement over relevance, and create dependency rather than building your skills. If Microsoft's AI assistant keeps inserting itself into workflows where you didn't ask for it, this framing may explain why.
Microsoft hasn't confirmed or denied the documents publicly. The company's official communications emphasize productivity improvements and responsible AI development. Both things can be true simultaneously in corporate messaging while pulling in opposite directions internally.
The more important question for daily users isn't whether this language reflects Microsoft's genuine intent - it almost certainly does. The question is whether the features being built serve your needs or theirs. Tracking what Copilot actually does with your attention over a month is more informative than reading the marketing materials.