Microsoft's secret document: the goal is to create AI dependence in users

A leaked internal Microsoft document suggests the company aims to make its AI chatbots habit-forming for users. Researchers and experts have previously warned about the risks of strong emotional attachments to artificial intelligence. The secret document raises questions about the ethics of AI design.

A leaked internal document from Microsoft reveals a troubling objective: the company wants to design its AI chatbots in a way that would create dependence in users. This disclosure has sparked broader debate about how large technology companies actually design AI assistants and where user interests fit in that process.

A familiar pattern in a new guise

Addictive design strategies are not entirely unfamiliar – social media platforms have long used similar tactics to keep users engaged with their applications for as long as possible. In the context of AI, however, this pattern takes on a fundamentally different dimension, as chatbots offer personal interaction that can create substantially stronger emotional bonds than conventional social media platforms.

Researchers and digital ethics experts have long warned that interactions with AI chatbots can develop into unexpectedly intense emotional relationships. This is particularly concerning for vulnerable user groups, such as lonely individuals, children, and those with mental health challenges, for whom a personal AI companion might seem extraordinarily appealing.

What exactly is this dependence?

The concept of dependence as used in Microsoft's document is ambiguous. On one hand, it could refer to cultivating productive habits – making users find the tool so useful that they return to it again and again. On the other hand, the word "dependence" suggests something fundamentally ethically problematic – an intention to manipulate user behaviour in a way that places commercial interests above user wellbeing.

Against a backdrop of increasing public scrutiny, AI-developing companies have come under mounting attention. Regulators in Europe and elsewhere are asking ever more pressing questions about how to ensure that AI systems serve users' interests, not merely those of their makers. Microsoft's secret document adds a new dimension to these discussions and raises the question of whether current regulations are sufficient to curb the risks posed by AI design ethics.