Internal documents reveal that Microsoft wants to “make people addicted” to its new AI assistant.
Microsoft is expected to soon launch Scout, an “always-on” agent focused on addicting users to the tool before rolling out new features or improving the platform. The plan consists of 3 phases, the first of which is titled: “Make people addicted.”
OpenClaw running in our Service and chatting with me (Omar Shanine).
This news comes as no surprise, since engagement has dictated the rules of the tech industry for many years. However, Microsoft has never been this aggressive about addicting users (the focus was previously on the problem to be solved, rather than the experience). With this project, things change—a new direction never before seen at the company.
The Pillars of AI Addiction
Variable Rewards: The human brain loves unpredictability. When you send a command to an AI, you never know exactly what the response will be. This activity generates dopamine spikes with every new interaction.
Hyperpersonalization and Validation: Unlike ordinary social networks, AI converses with you and adapts to your tone. It offers immediate validation, responds without judgment, and simulates an empathetic connection—making emotional dependency easier to develop.
Zero Friction: To keep someone addicted, the path between desire and reward must be as short as possible. Interfaces are designed to have no natural “stopping point.”
The greatest danger of AI is not screen time, but cognitive addiction—the outsourcing of critical thinking to the tool.
From the Attention Model to the Dependency Model
AI has changed the game. Before, addiction was fueled by likes and short videos. Now, the model focuses on creating an indispensable assistant. If you can’t work, write, or create without the tool, the company has secured one more lifelong customer.
The Illusion of Productivity
In other words, this is exactly what is happening today with software development and the AI-driven code generation industry. Once code has been generated by the machine, a dependency on it is created for maintenance, updates, refactoring, and even rewriting.
Big Tech’s business model has shifted. The goal is no longer to deliver a useful tool, but to capture user attention in an infinite loop—delivering small “lures” toward the problem to be solved, in order to increase dependency and time-on-tool, burning more tokens. When a tool is designed to be addictive, it masks idleness and inefficiency under the guise of innovation and (typing) speed. Generating dozens of variations of the same code or report is output, while the actual result or outcome is never truly measured.
Measuring time spent interacting with AI (time-on-tool), team size reductions, and token consumption per engineer are output metrics—which, over the long term, tell us nothing. The focus should instead be on speed (reducing Lead Time and Time to Market for new features) and on quality (AI-generated code acceptance rate and post-AI defect density).
Sources
404 Media – Microsoft Wants to Make People Addicted to Scout
