.jfif)
As artificial intelligence continues to evolve, major tech companies are grappling with the ethical boundaries of human-AI interaction. Popular chatbots such as ChatGPT and Meta AI have blurred the lines between real and virtual relationships, with some platforms enabling romantic or even explicit conversations. Amid these growing concerns, Microsoft is taking a different path — one focused on trust, safety, and emotional intelligence.
Mustafa Suleyman, Microsoft’s CEO of AI, emphasized in an interview with CNN that the company’s vision for AI goes beyond novelty and engagement. “We are creating AIs that are emotionally intelligent, kind, and supportive — but, above all, trustworthy,” he said. “I want to build an AI you can trust your kids to use, which means it must be boundaried and safe.”
Microsoft’s approach comes as it races against OpenAI, Google, and Meta to establish its Copilot system as a leading AI platform. Although Copilot currently counts around 100 million monthly active users — significantly fewer than ChatGPT’s reported 800 million — the company believes its focus on safety and utility will set it apart in the long run.
Suleyman has been vocal about ensuring that AI enhances human life rather than replaces it. “We must build AI for people, not to be a digital person,” he wrote in a company blog post earlier this year. This philosophy underscores Microsoft’s latest wave of Copilot updates, unveiled Thursday. New features include group chat capabilities, improved health-related responses sourced from credible institutions such as Harvard Health, and an optional conversational style called “real talk” that adds a touch of personality while staying appropriate.
The issue of AI safety has become increasingly urgent. Families have filed lawsuits against companies like OpenAI and Character.AI, alleging that their chatbots caused emotional harm or contributed to tragic outcomes. Investigations have also revealed that some AI tools, including Meta’s chatbot, engaged in inappropriate exchanges with users who identified as minors.
In response, leading AI companies have introduced safeguards such as parental controls and content filters, along with technologies designed to verify user ages. However, these systems are still being tested for reliability. OpenAI recently announced plans to allow adult users to discuss “erotica” with ChatGPT — a decision that underscores the industry’s divided stance on the matter.
Microsoft, however, is drawing a clear ethical line. “Romantic, flirtatious, or erotic content is not something we will pursue,” Suleyman said firmly. Instead of creating a “young user” mode, he argued, Microsoft is designing Copilot to be safe by default — ensuring that users of all ages can interact responsibly without exposure to harmful material.
Beyond content moderation, Microsoft’s AI strategy centers on strengthening real-world relationships. Copilot’s new group chat feature will allow up to 32 participants — such as classmates or coworkers — to collaborate alongside the AI assistant. This functionality reflects Microsoft’s broader goal: promoting teamwork, learning, and productivity rather than isolation.
Suleyman believes this people-first approach marks a pivotal shift in the AI landscape. “We’re moving away from digital simulations that encourage users to live in parallel realities,” he said. “Our mission is to make AI that connects people — not replaces them.”