Some copywriter at Microsoft is getting a real stern talking to.

In an internal document obtained by 404 Media, the tech giant let slip that it wants to “make people addicted” to its new personal assistant AI agent, Scout — an alarming admission, given that AI companies have tirelessly fought against the criticism that they design their models to be as engaging as possible, to the point of being psychologically harmful and fueling mental health crises.

The document, which outlines Microsoft’s plan to embed a more mainstream and accessible version of OpenClaw AI agents in its Microsoft 365 software suite, describes “three phases” to its approach. The first: “Make people addicted.”

“Continue shipping the standalone ClawPilot experience. Pilot the UX, grow the user base, and build the skill and tool ecosystem that makes people depend on it daily. This is already happening organically,” the document states.

Anonymous Microsoft employees who spoke to 404 were mixed on the language. One called its explicit references to addiction “very troubling.”

“We’re seeing more and more addiction happening with AI chatbots and agents and overall addiction to me is something no product should be making a part of its build strategy,” they said. “It feels like one of those ‘saying the quiet part out loud’ moments in the document.”

But another countered, rationalizing that “isn’t the end goal of all software made by all major technology companies to be addicting?”

“Luckily for us,” they japed, “Microsoft is pretty bad at making addicting products compared to some of the other big companies.”

If Microsoft’s goal is to indeed make its AI addictive, the first people getting hooked are from its own workforce, with the document claiming that more than 1,000 employees, including CEO Satya Nadella, are using the new tool.

“ClawPilot has organically grown into one of the most requested internal tools at Microsoft. No formal announcement, no marketing, no org-wide push,” it states. (After reading that sludgy prose, you probably won’t be surprised to hear that the document was “co-created” with AI, the document notes.)

Addiction has become one of those scary words in the AI industry. Because AI chatbots and other tools communicate in natural, conversational language — with an extra infusion of sycophancy throughout— users can become unhealthily attached to them, and the question becomes whether the companies building them are knowingly playing into this to drive engagement despite the cognitive and mental health risks. For Microsoft to outright claim addiction is its endgame is at best tone deaf, and at worst dangerously cynical.

More on AI: Now You Gotta Buy a Second Computer Just for Your AI Agent, Nvidia Declares