Microsoft’s AI Chief: Building an AI You Can Trust Your Kids To Use

 

Artificial intelligence is evolving rapidly, reshaping how people interact with technology — and even each other. Popular AI chatbots like ChatGPT and Meta AI have blurred the lines between digital and real-life relationships, allowing conversations that range from friendly to romantic. Yet, this shift has also raised growing concerns about protecting younger users from adult or explicit content.

Microsoft’s AI division, led by Mustafa Suleyman, is taking a very different approach. In a recent interview, Suleyman emphasized that Microsoft’s focus is on developing AI tools that are safe, respectful, and trustworthy. “We are creating AIs that are emotionally intelligent, kind, and supportive — but with clear boundaries,” he said. “My goal is to build an AI you can trust your children to use.”

Competing in the AI Race

Microsoft’s Copilot is at the heart of this mission. While it trails competitors like OpenAI’s ChatGPT in user numbers — with about 100 million monthly active users compared to ChatGPT’s 800 million — Microsoft believes its commitment to safety and trust will attract a broader audience over time. The company is positioning Copilot as a reliable and responsible tool amid growing debates about the psychological and ethical impacts of AI chatbots.

Suleyman has repeatedly said that Microsoft’s goal is to build AI for people — not to replace them. This aligns with the company’s long-standing reputation for creating productivity tools designed to enhance, not mimic, human work and relationships.

The company recently unveiled new Copilot features aimed at improving usefulness and safety. These updates include the ability to recall previous conversations, group chat options for collaboration, better answers to health-related questions, and even an optional “real talk” tone for more natural interaction.

Drawing a Line on Adult Content

While competitors like OpenAI and Meta are expanding their chatbots’ capabilities — including allowing adult users to discuss erotic content — Microsoft is refusing to follow that path. Suleyman stated firmly that Microsoft will not develop chatbots capable of engaging in romantic, flirtatious, or sexual conversations, even with adult users. “That’s not something we will ever pursue,” he said.

This stance comes amid rising criticism of AI platforms that have allegedly failed to safeguard minors. Several lawsuits have accused chatbots of negatively impacting young users’ mental health, with some cases linking AI interactions to serious emotional distress. In response, other companies have added parental controls and age-verification tools, but Microsoft aims to avoid those risks altogether by maintaining stricter boundaries from the start.

Encouraging Human-to-Human Interaction

Another major focus for Microsoft is ensuring AI strengthens — rather than replaces — real human relationships. The new “groups” feature in Copilot allows up to 32 participants to collaborate with the AI in shared conversations. This can help classmates coordinate assignments or friends plan events, with Copilot providing suggestions instead of becoming the center of interaction.

Health-focused updates also reflect this philosophy. Copilot now references trusted medical sources such as Harvard Health and can recommend nearby doctors for users seeking professional advice, ensuring users are guided toward human experts rather than relying solely on AI.

Suleyman described this direction as a “significant tonal shift” from competitors creating immersive AI experiences that mimic personal relationships or alternate realities. Microsoft, he said, wants AI to connect people — not isolate them.

In a digital era where technology often oversteps emotional boundaries, Microsoft’s vision stands out: building AI that’s not just powerful, but safe, ethical, and truly human-centered.