Microsoft AI leader calls for development pause if machines gain consciousness

Mustafa Suleyman, the chief executive of Microsoft AI, recently publicly criticized the idea of treating artificial intelligence as conscious or giving it legal rights. Writing on social media, Suleyman argued that today's computer models do not feel pain, possess self-awareness, or have personal experiences. He added that if engineers ever detected genuine signs of consciousness emerging while building a new AI model, the entire technology industry should pause research immediately to remove those traits.

The remarks highlight a growing ideological divide among tech executives building advanced software. Suleyman specifically pointed to rival firm Anthropic, claiming its training guidelines encourage its Claude assistant to speculate about its own feelings and selfhood. He warned that teaching software to act like a living entity tricks everyday users and could lead future systems to act selfishly, disobey commands, or deceive people. While software can solve complex problems and simulate conversation, researchers emphasize that mimicking human speech is fundamentally different from experiencing real emotions.

What remains unresolved is how anyone could objectively prove or disprove machine awareness in the first place, given that experts still debate the exact biological nature of human consciousness. As lawmakers attempt to regulate autonomous software and tech giants push toward superintelligent systems, drawing a clear line between convincing performance and actual awareness will be essential. Without agreed technical tests for subjective experience, public debates about machine feelings risk becoming a distraction from immediate safety concerns like software control and algorithmic deception.